diff options
Diffstat (limited to 'fs')
104 files changed, 3258 insertions, 2289 deletions
diff --git a/fs/binfmt_elf.c b/fs/binfmt_elf.c index a7a28110dc80..7f6aff3f72eb 100644 --- a/fs/binfmt_elf.c +++ b/fs/binfmt_elf.c @@ -605,28 +605,30 @@ static unsigned long load_elf_interp(struct elfhdr *interp_elf_ex, * Do the same thing for the memory mapping - between * elf_bss and last_bss is the bss section. */ - k = load_addr + eppnt->p_memsz + eppnt->p_vaddr; + k = load_addr + eppnt->p_vaddr + eppnt->p_memsz; if (k > last_bss) last_bss = k; } } + /* + * Now fill out the bss section: first pad the last page from + * the file up to the page boundary, and zero it from elf_bss + * up to the end of the page. + */ + if (padzero(elf_bss)) { + error = -EFAULT; + goto out; + } + /* + * Next, align both the file and mem bss up to the page size, + * since this is where elf_bss was just zeroed up to, and where + * last_bss will end after the vm_brk() below. + */ + elf_bss = ELF_PAGEALIGN(elf_bss); + last_bss = ELF_PAGEALIGN(last_bss); + /* Finally, if there is still more bss to allocate, do it. */ if (last_bss > elf_bss) { - /* - * Now fill out the bss section. First pad the last page up - * to the page boundary, and then perform a mmap to make sure - * that there are zero-mapped pages up to and including the - * last bss page. - */ - if (padzero(elf_bss)) { - error = -EFAULT; - goto out; - } - - /* What we have mapped so far */ - elf_bss = ELF_PAGESTART(elf_bss + ELF_MIN_ALIGN - 1); - - /* Map the last of the bss segment */ error = vm_brk(elf_bss, last_bss - elf_bss); if (error) goto out; diff --git a/fs/binfmt_em86.c b/fs/binfmt_em86.c index 490538536cb4..dd2d3f0cd55d 100644 --- a/fs/binfmt_em86.c +++ b/fs/binfmt_em86.c @@ -24,7 +24,8 @@ static int load_em86(struct linux_binprm *bprm) { - char *interp, *i_name, *i_arg; + const char *i_name, *i_arg; + char *interp; struct file * file; int retval; struct elfhdr elf_ex; diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index b2620d1f883f..443fcc402114 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -439,6 +439,8 @@ struct btrfs_space_info { struct list_head list; /* Protected by the spinlock 'lock'. */ struct list_head ro_bgs; + struct list_head priority_tickets; + struct list_head tickets; struct rw_semaphore groups_sem; /* for block groups in our same type */ @@ -2624,6 +2626,15 @@ enum btrfs_reserve_flush_enum { BTRFS_RESERVE_FLUSH_ALL, }; +enum btrfs_flush_state { + FLUSH_DELAYED_ITEMS_NR = 1, + FLUSH_DELAYED_ITEMS = 2, + FLUSH_DELALLOC = 3, + FLUSH_DELALLOC_WAIT = 4, + ALLOC_CHUNK = 5, + COMMIT_TRANS = 6, +}; + int btrfs_check_data_free_space(struct inode *inode, u64 start, u64 len); int btrfs_alloc_data_chunk_ondemand(struct inode *inode, u64 bytes); void btrfs_free_reserved_data_space(struct inode *inode, u64 start, u64 len); @@ -2661,8 +2672,8 @@ int btrfs_block_rsv_refill(struct btrfs_root *root, struct btrfs_block_rsv *block_rsv, u64 min_reserved, enum btrfs_reserve_flush_enum flush); int btrfs_block_rsv_migrate(struct btrfs_block_rsv *src_rsv, - struct btrfs_block_rsv *dst_rsv, - u64 num_bytes); + struct btrfs_block_rsv *dst_rsv, u64 num_bytes, + int update_size); int btrfs_cond_migrate_bytes(struct btrfs_fs_info *fs_info, struct btrfs_block_rsv *dest, u64 num_bytes, int min_factor); diff --git a/fs/btrfs/delayed-inode.c b/fs/btrfs/delayed-inode.c index d3aaabbfada0..dd3c040139a2 100644 --- a/fs/btrfs/delayed-inode.c +++ b/fs/btrfs/delayed-inode.c @@ -553,7 +553,7 @@ static int btrfs_delayed_item_reserve_metadata(struct btrfs_trans_handle *trans, dst_rsv = &root->fs_info->delayed_block_rsv; num_bytes = btrfs_calc_trans_metadata_size(root, 1); - ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes); + ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes, 1); if (!ret) { trace_btrfs_space_reservation(root->fs_info, "delayed_item", item->key.objectid, @@ -598,6 +598,29 @@ static int btrfs_delayed_inode_reserve_metadata( num_bytes = btrfs_calc_trans_metadata_size(root, 1); /* + * If our block_rsv is the delalloc block reserve then check and see if + * we have our extra reservation for updating the inode. If not fall + * through and try to reserve space quickly. + * + * We used to try and steal from the delalloc block rsv or the global + * reserve, but we'd steal a full reservation, which isn't kind. We are + * here through delalloc which means we've likely just cowed down close + * to the leaf that contains the inode, so we would steal less just + * doing the fallback inode update, so if we do end up having to steal + * from the global block rsv we hopefully only steal one or two blocks + * worth which is less likely to hurt us. + */ + if (src_rsv && src_rsv->type == BTRFS_BLOCK_RSV_DELALLOC) { + spin_lock(&BTRFS_I(inode)->lock); + if (test_and_clear_bit(BTRFS_INODE_DELALLOC_META_RESERVED, + &BTRFS_I(inode)->runtime_flags)) + release = true; + else + src_rsv = NULL; + spin_unlock(&BTRFS_I(inode)->lock); + } + + /* * btrfs_dirty_inode will update the inode under btrfs_join_transaction * which doesn't reserve space for speed. This is a problem since we * still need to reserve space for this update, so try to reserve the @@ -626,51 +649,10 @@ static int btrfs_delayed_inode_reserve_metadata( num_bytes, 1); } return ret; - } else if (src_rsv->type == BTRFS_BLOCK_RSV_DELALLOC) { - spin_lock(&BTRFS_I(inode)->lock); - if (test_and_clear_bit(BTRFS_INODE_DELALLOC_META_RESERVED, - &BTRFS_I(inode)->runtime_flags)) { - spin_unlock(&BTRFS_I(inode)->lock); - release = true; - goto migrate; - } - spin_unlock(&BTRFS_I(inode)->lock); - - /* Ok we didn't have space pre-reserved. This shouldn't happen - * too often but it can happen if we do delalloc to an existing - * inode which gets dirtied because of the time update, and then - * isn't touched again until after the transaction commits and - * then we try to write out the data. First try to be nice and - * reserve something strictly for us. If not be a pain and try - * to steal from the delalloc block rsv. - */ - ret = btrfs_block_rsv_add(root, dst_rsv, num_bytes, - BTRFS_RESERVE_NO_FLUSH); - if (!ret) - goto out; - - ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes); - if (!ret) - goto out; - - if (btrfs_test_opt(root, ENOSPC_DEBUG)) { - btrfs_debug(root->fs_info, - "block rsv migrate returned %d", ret); - WARN_ON(1); - } - /* - * Ok this is a problem, let's just steal from the global rsv - * since this really shouldn't happen that often. - */ - ret = btrfs_block_rsv_migrate(&root->fs_info->global_block_rsv, - dst_rsv, num_bytes); - goto out; } -migrate: - ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes); + ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes, 1); -out: /* * Migrate only takes a reservation, it doesn't touch the size of the * block_rsv. This is to simplify people who don't normally have things diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index b480fd555774..e9376b1657e2 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -111,6 +111,16 @@ static int block_rsv_use_bytes(struct btrfs_block_rsv *block_rsv, u64 num_bytes); int btrfs_pin_extent(struct btrfs_root *root, u64 bytenr, u64 num_bytes, int reserved); +static int __reserve_metadata_bytes(struct btrfs_root *root, + struct btrfs_space_info *space_info, + u64 orig_bytes, + enum btrfs_reserve_flush_enum flush); +static void space_info_add_new_bytes(struct btrfs_fs_info *fs_info, + struct btrfs_space_info *space_info, + u64 num_bytes); +static void space_info_add_old_bytes(struct btrfs_fs_info *fs_info, + struct btrfs_space_info *space_info, + u64 num_bytes); static noinline int block_group_cache_done(struct btrfs_block_group_cache *cache) @@ -3913,6 +3923,7 @@ static const char *alloc_name(u64 flags) static int update_space_info(struct btrfs_fs_info *info, u64 flags, u64 total_bytes, u64 bytes_used, + u64 bytes_readonly, struct btrfs_space_info **space_info) { struct btrfs_space_info *found; @@ -3933,8 +3944,11 @@ static int update_space_info(struct btrfs_fs_info *info, u64 flags, found->disk_total += total_bytes * factor; found->bytes_used += bytes_used; found->disk_used += bytes_used * factor; + found->bytes_readonly += bytes_readonly; if (total_bytes > 0) found->full = 0; + space_info_add_new_bytes(info, found, total_bytes - + bytes_used - bytes_readonly); spin_unlock(&found->lock); *space_info = found; return 0; @@ -3960,7 +3974,7 @@ static int update_space_info(struct btrfs_fs_info *info, u64 flags, found->disk_used = bytes_used * factor; found->bytes_pinned = 0; found->bytes_reserved = 0; - found->bytes_readonly = 0; + found->bytes_readonly = bytes_readonly; found->bytes_may_use = 0; found->full = 0; found->max_extent_size = 0; @@ -3969,6 +3983,8 @@ static int update_space_info(struct btrfs_fs_info *info, u64 flags, found->flush = 0; init_waitqueue_head(&found->wait); INIT_LIST_HEAD(&found->ro_bgs); + INIT_LIST_HEAD(&found->tickets); + INIT_LIST_HEAD(&found->priority_tickets); ret = kobject_init_and_add(&found->kobj, &space_info_ktype, info->space_info_kobj, "%s", @@ -4470,7 +4486,7 @@ static int do_chunk_alloc(struct btrfs_trans_handle *trans, space_info = __find_space_info(extent_root->fs_info, flags); if (!space_info) { ret = update_space_info(extent_root->fs_info, flags, - 0, 0, &space_info); + 0, 0, 0, &space_info); BUG_ON(ret); /* -ENOMEM */ } BUG_ON(!space_info); /* Logic error */ @@ -4582,12 +4598,19 @@ static int can_overcommit(struct btrfs_root *root, struct btrfs_space_info *space_info, u64 bytes, enum btrfs_reserve_flush_enum flush) { - struct btrfs_block_rsv *global_rsv = &root->fs_info->global_block_rsv; - u64 profile = btrfs_get_alloc_profile(root, 0); + struct btrfs_block_rsv *global_rsv; + u64 profile; u64 space_size; u64 avail; u64 used; + /* Don't overcommit when in mixed mode. */ + if (space_info->flags & BTRFS_BLOCK_GROUP_DATA) + return 0; + + BUG_ON(root->fs_info == NULL); + global_rsv = &root->fs_info->global_block_rsv; + profile = btrfs_get_alloc_profile(root, 0); used = space_info->bytes_used + space_info->bytes_reserved + space_info->bytes_pinned + space_info->bytes_readonly; @@ -4739,6 +4762,11 @@ skip_async: spin_unlock(&space_info->lock); break; } + if (list_empty(&space_info->tickets) && + list_empty(&space_info->priority_tickets)) { + spin_unlock(&space_info->lock); + break; + } spin_unlock(&space_info->lock); loops++; @@ -4807,13 +4835,11 @@ commit: return btrfs_commit_transaction(trans, root); } -enum flush_state { - FLUSH_DELAYED_ITEMS_NR = 1, - FLUSH_DELAYED_ITEMS = 2, - FLUSH_DELALLOC = 3, - FLUSH_DELALLOC_WAIT = 4, - ALLOC_CHUNK = 5, - COMMIT_TRANS = 6, +struct reserve_ticket { + u64 bytes; + int error; + struct list_head list; + wait_queue_head_t wait; }; static int flush_space(struct btrfs_root *root, @@ -4866,6 +4892,8 @@ static int flush_space(struct btrfs_root *root, break; } + trace_btrfs_flush_space(root->fs_info, space_info->flags, num_bytes, + orig_bytes, state, ret); return ret; } @@ -4873,17 +4901,22 @@ static inline u64 btrfs_calc_reclaim_metadata_size(struct btrfs_root *root, struct btrfs_space_info *space_info) { + struct reserve_ticket *ticket; u64 used; u64 expected; - u64 to_reclaim; + u64 to_reclaim = 0; to_reclaim = min_t(u64, num_online_cpus() * SZ_1M, SZ_16M); - spin_lock(&space_info->lock); if (can_overcommit(root, space_info, to_reclaim, - BTRFS_RESERVE_FLUSH_ALL)) { - to_reclaim = 0; - goto out; - } + BTRFS_RESERVE_FLUSH_ALL)) + return 0; + + list_for_each_entry(ticket, &space_info->tickets, list) + to_reclaim += ticket->bytes; + list_for_each_entry(ticket, &space_info->priority_tickets, list) + to_reclaim += ticket->bytes; + if (to_reclaim) + return to_reclaim; used = space_info->bytes_used + space_info->bytes_reserved + space_info->bytes_pinned + space_info->bytes_readonly + @@ -4899,14 +4932,11 @@ btrfs_calc_reclaim_metadata_size(struct btrfs_root *root, to_reclaim = 0; to_reclaim = min(to_reclaim, space_info->bytes_may_use + space_info->bytes_reserved); -out: - spin_unlock(&space_info->lock); - return to_reclaim; } static inline int need_do_async_reclaim(struct btrfs_space_info *space_info, - struct btrfs_fs_info *fs_info, u64 used) + struct btrfs_root *root, u64 used) { u64 thresh = div_factor_fine(space_info->total_bytes, 98); @@ -4914,73 +4944,177 @@ static inline int need_do_async_reclaim(struct btrfs_space_info *space_info, if ((space_info->bytes_used + space_info->bytes_reserved) >= thresh) return 0; - return (used >= thresh && !btrfs_fs_closing(fs_info) && - !test_bit(BTRFS_FS_STATE_REMOUNTING, &fs_info->fs_state)); + if (!btrfs_calc_reclaim_metadata_size(root, space_info)) + return 0; + + return (used >= thresh && !btrfs_fs_closing(root->fs_info) && + !test_bit(BTRFS_FS_STATE_REMOUNTING, + &root->fs_info->fs_state)); } -static int btrfs_need_do_async_reclaim(struct btrfs_space_info *space_info, - struct btrfs_fs_info *fs_info, - int flush_state) +static void wake_all_tickets(struct list_head *head) { - u64 used; - - spin_lock(&space_info->lock); - /* - * We run out of space and have not got any free space via flush_space, - * so don't bother doing async reclaim. - */ - if (flush_state > COMMIT_TRANS && space_info->full) { - spin_unlock(&space_info->lock); - return 0; - } + struct reserve_ticket *ticket; - used = space_info->bytes_used + space_info->bytes_reserved + - space_info->bytes_pinned + space_info->bytes_readonly + - space_info->bytes_may_use; - if (need_do_async_reclaim(space_info, fs_info, used)) { - spin_unlock(&space_info->lock); - return 1; + while (!list_empty(head)) { + ticket = list_first_entry(head, struct reserve_ticket, list); + list_del_init(&ticket->list); + ticket->error = -ENOSPC; + wake_up(&ticket->wait); } - spin_unlock(&space_info->lock); - - return 0; } +/* + * This is for normal flushers, we can wait all goddamned day if we want to. We + * will loop and continuously try to flush as long as we are making progress. + * We count progress as clearing off tickets each time we have to loop. + */ static void btrfs_async_reclaim_metadata_space(struct work_struct *work) { + struct reserve_ticket *last_ticket = NULL; struct btrfs_fs_info *fs_info; struct btrfs_space_info *space_info; u64 to_reclaim; int flush_state; + int commit_cycles = 0; fs_info = container_of(work, struct btrfs_fs_info, async_reclaim_work); space_info = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_METADATA); + spin_lock(&space_info->lock); to_reclaim = btrfs_calc_reclaim_metadata_size(fs_info->fs_root, space_info); - if (!to_reclaim) + if (!to_reclaim) { + space_info->flush = 0; + spin_unlock(&space_info->lock); return; + } + last_ticket = list_first_entry(&space_info->tickets, + struct reserve_ticket, list); + spin_unlock(&space_info->lock); flush_state = FLUSH_DELAYED_ITEMS_NR; do { + struct reserve_ticket *ticket; + int ret; + + ret = flush_space(fs_info->fs_root, space_info, to_reclaim, + to_reclaim, flush_state); + spin_lock(&space_info->lock); + if (list_empty(&space_info->tickets)) { + space_info->flush = 0; + spin_unlock(&space_info->lock); + return; + } + to_reclaim = btrfs_calc_reclaim_metadata_size(fs_info->fs_root, + space_info); + ticket = list_first_entry(&space_info->tickets, + struct reserve_ticket, list); + if (last_ticket == ticket) { + flush_state++; + } else { + last_ticket = ticket; + flush_state = FLUSH_DELAYED_ITEMS_NR; + if (commit_cycles) + commit_cycles--; + } + + if (flush_state > COMMIT_TRANS) { + commit_cycles++; + if (commit_cycles > 2) { + wake_all_tickets(&space_info->tickets); + space_info->flush = 0; + } else { + flush_state = FLUSH_DELAYED_ITEMS_NR; + } + } + spin_unlock(&space_info->lock); + } while (flush_state <= COMMIT_TRANS); +} + +void btrfs_init_async_reclaim_work(struct work_struct *work) +{ + INIT_WORK(work, btrfs_async_reclaim_metadata_space); +} + +static void priority_reclaim_metadata_space(struct btrfs_fs_info *fs_info, + struct btrfs_space_info *space_info, + struct reserve_ticket *ticket) +{ + u64 to_reclaim; + int flush_state = FLUSH_DELAYED_ITEMS_NR; + + spin_lock(&space_info->lock); + to_reclaim = btrfs_calc_reclaim_metadata_size(fs_info->fs_root, + space_info); + if (!to_reclaim) { + spin_unlock(&space_info->lock); + return; + } + spin_unlock(&space_info->lock); + + do { flush_space(fs_info->fs_root, space_info, to_reclaim, to_reclaim, flush_state); flush_state++; - if (!btrfs_need_do_async_reclaim(space_info, fs_info, - flush_state)) + spin_lock(&space_info->lock); + if (ticket->bytes == 0) { + spin_unlock(&space_info->lock); return; + } + spin_unlock(&space_info->lock); + + /* + * Priority flushers can't wait on delalloc without + * deadlocking. + */ + if (flush_state == FLUSH_DELALLOC || + flush_state == FLUSH_DELALLOC_WAIT) + flush_state = ALLOC_CHUNK; } while (flush_state < COMMIT_TRANS); } -void btrfs_init_async_reclaim_work(struct work_struct *work) +static int wait_reserve_ticket(struct btrfs_fs_info *fs_info, + struct btrfs_space_info *space_info, + struct reserve_ticket *ticket, u64 orig_bytes) + { - INIT_WORK(work, btrfs_async_reclaim_metadata_space); + DEFINE_WAIT(wait); + int ret = 0; + + spin_lock(&space_info->lock); + while (ticket->bytes > 0 && ticket->error == 0) { + ret = prepare_to_wait_event(&ticket->wait, &wait, TASK_KILLABLE); + if (ret) { + ret = -EINTR; + break; + } + spin_unlock(&space_info->lock); + + schedule(); + + finish_wait(&ticket->wait, &wait); + spin_lock(&space_info->lock); + } + if (!ret) + ret = ticket->error; + if (!list_empty(&ticket->list)) + list_del_init(&ticket->list); + if (ticket->bytes && ticket->bytes < orig_bytes) { + u64 num_bytes = orig_bytes - ticket->bytes; + space_info->bytes_may_use -= num_bytes; + trace_btrfs_space_reservation(fs_info, "space_info", + space_info->flags, num_bytes, 0); + } + spin_unlock(&space_info->lock); + + return ret; } /** * reserve_metadata_bytes - try to reserve bytes from the block_rsv's space * @root - the root we're allocating for - * @block_rsv - the block_rsv we're allocating for + * @space_info - the space info we want to allocate from * @orig_bytes - the number of bytes we want * @flush - whether or not we can flush to make our reservation * @@ -4991,81 +5125,36 @@ void btrfs_init_async_reclaim_work(struct work_struct *work) * regain reservations will be made and this will fail if there is not enough * space already. */ -static int reserve_metadata_bytes(struct btrfs_root *root, - struct btrfs_block_rsv *block_rsv, - u64 orig_bytes, - enum btrfs_reserve_flush_enum flush) +static int __reserve_metadata_bytes(struct btrfs_root *root, + struct btrfs_space_info *space_info, + u64 orig_bytes, + enum btrfs_reserve_flush_enum flush) { - struct btrfs_space_info *space_info = block_rsv->space_info; + struct reserve_ticket ticket; u64 used; - u64 num_bytes = orig_bytes; - int flush_state = FLUSH_DELAYED_ITEMS_NR; int ret = 0; - bool flushing = false; -again: - ret = 0; - spin_lock(&space_info->lock); - /* - * We only want to wait if somebody other than us is flushing and we - * are actually allowed to flush all things. - */ - while (flush == BTRFS_RESERVE_FLUSH_ALL && !flushing && - space_info->flush) { - spin_unlock(&space_info->lock); - /* - * If we have a trans handle we can't wait because the flusher - * may have to commit the transaction, which would mean we would - * deadlock since we are waiting for the flusher to finish, but - * hold the current transaction open. - */ - if (current->journal_info) - return -EAGAIN; - ret = wait_event_killable(space_info->wait, !space_info->flush); - /* Must have been killed, return */ - if (ret) - return -EINTR; - - spin_lock(&space_info->lock); - } + ASSERT(orig_bytes); + ASSERT(!current->journal_info || flush != BTRFS_RESERVE_FLUSH_ALL); + spin_lock(&space_info->lock); ret = -ENOSPC; used = space_info->bytes_used + space_info->bytes_reserved + space_info->bytes_pinned + space_info->bytes_readonly + space_info->bytes_may_use; /* - * The idea here is that we've not already over-reserved the block group - * then we can go ahead and save our reservation first and then start - * flushing if we need to. Otherwise if we've already overcommitted - * lets start flushing stuff first and then come back and try to make - * our reservation. + * If we have enough space then hooray, make our reservation and carry + * on. If not see if we can overcommit, and if we can, hooray carry on. + * If not things get more complicated. */ - if (used <= space_info->total_bytes) { - if (used + orig_bytes <= space_info->total_bytes) { - space_info->bytes_may_use += orig_bytes; - trace_btrfs_space_reservation(root->fs_info, - "space_info", space_info->flags, orig_bytes, 1); - ret = 0; - } else { - /* - * Ok set num_bytes to orig_bytes since we aren't - * overocmmitted, this way we only try and reclaim what - * we need. - */ - num_bytes = orig_bytes; - } - } else { - /* - * Ok we're over committed, set num_bytes to the overcommitted - * amount plus the amount of bytes that we need for this - * reservation. - */ - num_bytes = used - space_info->total_bytes + - (orig_bytes * 2); - } - - if (ret && can_overcommit(root, space_info, orig_bytes, flush)) { + if (used + orig_bytes <= space_info->total_bytes) { + space_info->bytes_may_use += orig_bytes; + trace_btrfs_space_reservation(root->fs_info, "space_info", + space_info->flags, orig_bytes, + 1); + ret = 0; + } else if (can_overcommit(root, space_info, orig_bytes, flush)) { space_info->bytes_may_use += orig_bytes; trace_btrfs_space_reservation(root->fs_info, "space_info", space_info->flags, orig_bytes, @@ -5074,16 +5163,31 @@ again: } /* - * Couldn't make our reservation, save our place so while we're trying - * to reclaim space we can actually use it instead of somebody else - * stealing it from us. + * If we couldn't make a reservation then setup our reservation ticket + * and kick the async worker if it's not already running. * - * We make the other tasks wait for the flush only when we can flush - * all things. + * If we are a priority flusher then we just need to add our ticket to + * the list and we will do our own flushing further down. */ if (ret && flush != BTRFS_RESERVE_NO_FLUSH) { - flushing = true; - space_info->flush = 1; + ticket.bytes = orig_bytes; + ticket.error = 0; + init_waitqueue_head(&ticket.wait); + if (flush == BTRFS_RESERVE_FLUSH_ALL) { + list_add_tail(&ticket.list, &space_info->tickets); + if (!space_info->flush) { + space_info->flush = 1; + trace_btrfs_trigger_flush(root->fs_info, + space_info->flags, + orig_bytes, flush, + "enospc"); + queue_work(system_unbound_wq, + &root->fs_info->async_reclaim_work); + } + } else { + list_add_tail(&ticket.list, + &space_info->priority_tickets); + } } else if (!ret && space_info->flags & BTRFS_BLOCK_GROUP_METADATA) { used += orig_bytes; /* @@ -5092,39 +5196,67 @@ again: * the async reclaim as we will panic. */ if (!root->fs_info->log_root_recovering && - need_do_async_reclaim(space_info, root->fs_info, used) && - !work_busy(&root->fs_info->async_reclaim_work)) + need_do_async_reclaim(space_info, root, used) && + !work_busy(&root->fs_info->async_reclaim_work)) { + trace_btrfs_trigger_flush(root->fs_info, + space_info->flags, + orig_bytes, flush, + "preempt"); queue_work(system_unbound_wq, &root->fs_info->async_reclaim_work); + } } spin_unlock(&space_info->lock); - if (!ret || flush == BTRFS_RESERVE_NO_FLUSH) - goto out; + return ret; - ret = flush_space(root, space_info, num_bytes, orig_bytes, - flush_state); - flush_state++; + if (flush == BTRFS_RESERVE_FLUSH_ALL) + return wait_reserve_ticket(root->fs_info, space_info, &ticket, + orig_bytes); - /* - * If we are FLUSH_LIMIT, we can not flush delalloc, or the deadlock - * would happen. So skip delalloc flush. - */ - if (flush == BTRFS_RESERVE_FLUSH_LIMIT && - (flush_state == FLUSH_DELALLOC || - flush_state == FLUSH_DELALLOC_WAIT)) - flush_state = ALLOC_CHUNK; + ret = 0; + priority_reclaim_metadata_space(root->fs_info, space_info, &ticket); + spin_lock(&space_info->lock); + if (ticket.bytes) { + if (ticket.bytes < orig_bytes) { + u64 num_bytes = orig_bytes - ticket.bytes; + space_info->bytes_may_use -= num_bytes; + trace_btrfs_space_reservation(root->fs_info, + "space_info", space_info->flags, + num_bytes, 0); - if (!ret) - goto again; - else if (flush == BTRFS_RESERVE_FLUSH_LIMIT && - flush_state < COMMIT_TRANS) - goto again; - else if (flush == BTRFS_RESERVE_FLUSH_ALL && - flush_state <= COMMIT_TRANS) - goto again; + } + list_del_init(&ticket.list); + ret = -ENOSPC; + } + spin_unlock(&space_info->lock); + ASSERT(list_empty(&ticket.list)); + return ret; +} -out: +/** + * reserve_metadata_bytes - try to reserve bytes from the block_rsv's space + * @root - the root we're allocating for + * @block_rsv - the block_rsv we're allocating for + * @orig_bytes - the number of bytes we want + * @flush - whether or not we can flush to make our reservation + * + * This will reserve orgi_bytes number of bytes from the space info associated + * with the block_rsv. If there is not enough space it will make an attempt to + * flush out space to make room. It will do this by flushing delalloc if + * possible or committing the transaction. If flush is 0 then no attempts to + * regain reservations will be made and this will fail if there is not enough + * space already. + */ +static int reserve_metadata_bytes(struct btrfs_root *root, + struct btrfs_block_rsv *block_rsv, + u64 orig_bytes, + enum btrfs_reserve_flush_enum flush) +{ + int ret; + + ret = __reserve_metadata_bytes(root, block_rsv->space_info, orig_bytes, + flush); if (ret == -ENOSPC && unlikely(root->orphan_cleanup_state == ORPHAN_CLEANUP_STARTED)) { struct btrfs_block_rsv *global_rsv = @@ -5137,13 +5269,8 @@ out: if (ret == -ENOSPC) trace_btrfs_space_reservation(root->fs_info, "space_info:enospc", - space_info->flags, orig_bytes, 1); - if (flushing) { - spin_lock(&space_info->lock); - space_info->flush = 0; - wake_up_all(&space_info->wait); - spin_unlock(&space_info->lock); - } + block_rsv->space_info->flags, + orig_bytes, 1); return ret; } @@ -5219,6 +5346,108 @@ int btrfs_cond_migrate_bytes(struct btrfs_fs_info *fs_info, return 0; } +/* + * This is for space we already have accounted in space_info->bytes_may_use, so + * basically when we're returning space from block_rsv's. + */ +static void space_info_add_old_bytes(struct btrfs_fs_info *fs_info, + struct btrfs_space_info *space_info, + u64 num_bytes) +{ + struct reserve_ticket *ticket; + struct list_head *head; + u64 used; + enum btrfs_reserve_flush_enum flush = BTRFS_RESERVE_NO_FLUSH; + bool check_overcommit = false; + + spin_lock(&space_info->lock); + head = &space_info->priority_tickets; + + /* + * If we are over our limit then we need to check and see if we can + * overcommit, and if we can't then we just need to free up our space + * and not satisfy any requests. + */ + used = space_info->bytes_used + space_info->bytes_reserved + + space_info->bytes_pinned + space_info->bytes_readonly + + space_info->bytes_may_use; + if (used - num_bytes >= space_info->total_bytes) + check_overcommit = true; +again: + while (!list_empty(head) && num_bytes) { + ticket = list_first_entry(head, struct reserve_ticket, + list); + /* + * We use 0 bytes because this space is already reserved, so + * adding the ticket space would be a double count. + */ + if (check_overcommit && + !can_overcommit(fs_info->extent_root, space_info, 0, + flush)) + break; + if (num_bytes >= ticket->bytes) { + list_del_init(&ticket->list); + num_bytes -= ticket->bytes; + ticket->bytes = 0; + wake_up(&ticket->wait); + } else { + ticket->bytes -= num_bytes; + num_bytes = 0; + } + } + + if (num_bytes && head == &space_info->priority_tickets) { + head = &space_info->tickets; + flush = BTRFS_RESERVE_FLUSH_ALL; + goto again; + } + space_info->bytes_may_use -= num_bytes; + trace_btrfs_space_reservation(fs_info, "space_info", + space_info->flags, num_bytes, 0); + spin_unlock(&space_info->lock); +} + +/* + * This is for newly allocated space that isn't accounted in + * space_info->bytes_may_use yet. So if we allocate a chunk or unpin an extent + * we use this helper. + */ +static void space_info_add_new_bytes(struct btrfs_fs_info *fs_info, + struct btrfs_space_info *space_info, + u64 num_bytes) +{ + struct reserve_ticket *ticket; + struct list_head *head = &space_info->priority_tickets; + +again: + while (!list_empty(head) && num_bytes) { + ticket = list_first_entry(head, struct reserve_ticket, + list); + if (num_bytes >= ticket->bytes) { + trace_btrfs_space_reservation(fs_info, "space_info", + space_info->flags, + ticket->bytes, 1); + list_del_init(&ticket->list); + num_bytes -= ticket->bytes; + space_info->bytes_may_use += ticket->bytes; + ticket->bytes = 0; + wake_up(&ticket->wait); + } else { + trace_btrfs_space_reservation(fs_info, "space_info", + space_info->flags, + num_bytes, 1); + space_info->bytes_may_use += num_bytes; + ticket->bytes -= num_bytes; + num_bytes = 0; + } + } + + if (num_bytes && head == &space_info->priority_tickets) { + head = &space_info->tickets; + goto again; + } +} + static void block_rsv_release_bytes(struct btrfs_fs_info *fs_info, struct btrfs_block_rsv *block_rsv, struct btrfs_block_rsv *dest, u64 num_bytes) @@ -5253,18 +5482,15 @@ static void block_rsv_release_bytes(struct btrfs_fs_info *fs_info, } spin_unlock(&dest->lock); } - if (num_bytes) { - spin_lock(&space_info->lock); - space_info->bytes_may_use -= num_bytes; - trace_btrfs_space_reservation(fs_info, "space_info", - space_info->flags, num_bytes, 0); - spin_unlock(&space_info->lock); - } + if (num_bytes) + space_info_add_old_bytes(fs_info, space_info, + num_bytes); } } -static int block_rsv_migrate_bytes(struct btrfs_block_rsv *src, - struct btrfs_block_rsv *dst, u64 num_bytes) +int btrfs_block_rsv_migrate(struct btrfs_block_rsv *src, + struct btrfs_block_rsv *dst, u64 num_bytes, + int update_size) { int ret; @@ -5272,7 +5498,7 @@ static int block_rsv_migrate_bytes(struct btrfs_block_rsv *src, if (ret) return ret; - block_rsv_add_bytes(dst, num_bytes, 1); + block_rsv_add_bytes(dst, num_bytes, update_size); return 0; } @@ -5379,13 +5605,6 @@ int btrfs_block_rsv_refill(struct btrfs_root *root, return ret; } -int btrfs_block_rsv_migrate(struct btrfs_block_rsv *src_rsv, - struct btrfs_block_rsv *dst_rsv, - u64 num_bytes) -{ - return block_rsv_migrate_bytes(src_rsv, dst_rsv, num_bytes); -} - void btrfs_block_rsv_release(struct btrfs_root *root, struct btrfs_block_rsv *block_rsv, u64 num_bytes) @@ -5398,48 +5617,21 @@ void btrfs_block_rsv_release(struct btrfs_root *root, num_bytes); } -/* - * helper to calculate size of global block reservation. - * the desired value is sum of space used by extent tree, - * checksum tree and root tree - */ -static u64 calc_global_metadata_size(struct btrfs_fs_info *fs_info) -{ - struct btrfs_space_info *sinfo; - u64 num_bytes; - u64 meta_used; - u64 data_used; - int csum_size = btrfs_super_csum_size(fs_info->super_copy); - - sinfo = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_DATA); - spin_lock(&sinfo->lock); - data_used = sinfo->bytes_used; - spin_unlock(&sinfo->lock); - - sinfo = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_METADATA); - spin_lock(&sinfo->lock); - if (sinfo->flags & BTRFS_BLOCK_GROUP_DATA) - data_used = 0; - meta_used = sinfo->bytes_used; - spin_unlock(&sinfo->lock); - - num_bytes = (data_used >> fs_info->sb->s_blocksize_bits) * - csum_size * 2; - num_bytes += div_u64(data_used + meta_used, 50); - - if (num_bytes * 3 > meta_used) - num_bytes = div_u64(meta_used, 3); - - return ALIGN(num_bytes, fs_info->extent_root->nodesize << 10); -} - static void update_global_block_rsv(struct btrfs_fs_info *fs_info) { struct btrfs_block_rsv *block_rsv = &fs_info->global_block_rsv; struct btrfs_space_info *sinfo = block_rsv->space_info; u64 num_bytes; - num_bytes = calc_global_metadata_size(fs_info); + /* + * The global block rsv is based on the size of the extent tree, the + * checksum tree and the root tree. If the fs is empty we want to set + * it to a minimal amount for safety. + */ + num_bytes = btrfs_root_used(&fs_info->extent_root->root_item) + + btrfs_root_used(&fs_info->csum_root->root_item) + + btrfs_root_used(&fs_info->tree_root->root_item); + num_bytes = max_t(u64, num_bytes, SZ_16M); spin_lock(&sinfo->lock); spin_lock(&block_rsv->lock); @@ -5554,7 +5746,13 @@ int btrfs_orphan_reserve_metadata(struct btrfs_trans_handle *trans, struct inode *inode) { struct btrfs_root *root = BTRFS_I(inode)->root; - struct btrfs_block_rsv *src_rsv = get_block_rsv(trans, root); + /* + * We always use trans->block_rsv here as we will have reserved space + * for our orphan when starting the transaction, using get_block_rsv() + * here will sometimes make us choose the wrong block rsv as we could be + * doing a reloc inode for a non refcounted root. + */ + struct btrfs_block_rsv *src_rsv = trans->block_rsv; struct btrfs_block_rsv *dst_rsv = root->orphan_block_rsv; /* @@ -5565,7 +5763,7 @@ int btrfs_orphan_reserve_metadata(struct btrfs_trans_handle *trans, u64 num_bytes = btrfs_calc_trans_metadata_size(root, 1); trace_btrfs_space_reservation(root->fs_info, "orphan", btrfs_ino(inode), num_bytes, 1); - return block_rsv_migrate_bytes(src_rsv, dst_rsv, num_bytes); + return btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes, 1); } void btrfs_orphan_release_metadata(struct inode *inode) @@ -5620,7 +5818,7 @@ int btrfs_subvolume_reserve_metadata(struct btrfs_root *root, BTRFS_RESERVE_FLUSH_ALL); if (ret == -ENOSPC && use_global_rsv) - ret = btrfs_block_rsv_migrate(global_rsv, rsv, num_bytes); + ret = btrfs_block_rsv_migrate(global_rsv, rsv, num_bytes, 1); if (ret && *qgroup_reserved) btrfs_qgroup_free_meta(root, *qgroup_reserved); @@ -5730,21 +5928,26 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes) u64 to_reserve = 0; u64 csum_bytes; unsigned nr_extents = 0; - int extra_reserve = 0; enum btrfs_reserve_flush_enum flush = BTRFS_RESERVE_FLUSH_ALL; int ret = 0; bool delalloc_lock = true; u64 to_free = 0; unsigned dropped; + bool release_extra = false; /* If we are a free space inode we need to not flush since we will be in * the middle of a transaction commit. We also don't need the delalloc * mutex since we won't race with anybody. We need this mostly to make * lockdep shut its filthy mouth. + * + * If we have a transaction open (can happen if we call truncate_block + * from truncate), then we need FLUSH_LIMIT so we don't deadlock. */ if (btrfs_is_free_space_inode(inode)) { flush = BTRFS_RESERVE_NO_FLUSH; delalloc_lock = false; + } else if (current->journal_info) { + flush = BTRFS_RESERVE_FLUSH_LIMIT; } if (flush != BTRFS_RESERVE_NO_FLUSH && @@ -5761,24 +5964,15 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes) BTRFS_MAX_EXTENT_SIZE - 1, BTRFS_MAX_EXTENT_SIZE); BTRFS_I(inode)->outstanding_extents += nr_extents; - nr_extents = 0; + nr_extents = 0; if (BTRFS_I(inode)->outstanding_extents > BTRFS_I(inode)->reserved_extents) - nr_extents = BTRFS_I(inode)->outstanding_extents - + nr_extents += BTRFS_I(inode)->outstanding_extents - BTRFS_I(inode)->reserved_extents; - /* - * Add an item to reserve for updating the inode when we complete the - * delalloc io. - */ - if (!test_bit(BTRFS_INODE_DELALLOC_META_RESERVED, - &BTRFS_I(inode)->runtime_flags)) { - nr_extents++; - extra_reserve = 1; - } - - to_reserve = btrfs_calc_trans_metadata_size(root, nr_extents); + /* We always want to reserve a slot for updating the inode. */ + to_reserve = btrfs_calc_trans_metadata_size(root, nr_extents + 1); to_reserve += calc_csum_metadata_size(inode, num_bytes, 1); csum_bytes = BTRFS_I(inode)->csum_bytes; spin_unlock(&BTRFS_I(inode)->lock); @@ -5790,17 +5984,17 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes) goto out_fail; } - ret = reserve_metadata_bytes(root, block_rsv, to_reserve, flush); + ret = btrfs_block_rsv_add(root, block_rsv, to_reserve, flush); if (unlikely(ret)) { btrfs_qgroup_free_meta(root, nr_extents * root->nodesize); goto out_fail; } spin_lock(&BTRFS_I(inode)->lock); - if (extra_reserve) { - set_bit(BTRFS_INODE_DELALLOC_META_RESERVED, - &BTRFS_I(inode)->runtime_flags); - nr_extents--; + if (test_and_set_bit(BTRFS_INODE_DELALLOC_META_RESERVED, + &BTRFS_I(inode)->runtime_flags)) { + to_reserve -= btrfs_calc_trans_metadata_size(root, 1); + release_extra = true; } BTRFS_I(inode)->reserved_extents += nr_extents; spin_unlock(&BTRFS_I(inode)->lock); @@ -5811,8 +6005,10 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes) if (to_reserve) trace_btrfs_space_reservation(root->fs_info, "delalloc", btrfs_ino(inode), to_reserve, 1); - block_rsv_add_bytes(block_rsv, to_reserve, 1); - + if (release_extra) + btrfs_block_rsv_release(root, block_rsv, + btrfs_calc_trans_metadata_size(root, + 1)); return 0; out_fail: @@ -6044,6 +6240,9 @@ static int update_block_group(struct btrfs_trans_handle *trans, spin_unlock(&cache->lock); spin_unlock(&cache->space_info->lock); + trace_btrfs_space_reservation(root->fs_info, "pinned", + cache->space_info->flags, + num_bytes, 1); set_extent_dirty(info->pinned_extents, bytenr, bytenr + num_bytes - 1, GFP_NOFS | __GFP_NOFAIL); @@ -6118,10 +6317,10 @@ static int pin_down_extent(struct btrfs_root *root, spin_unlock(&cache->lock); spin_unlock(&cache->space_info->lock); + trace_btrfs_space_reservation(root->fs_info, "pinned", + cache->space_info->flags, num_bytes, 1); set_extent_dirty(root->fs_info->pinned_extents, bytenr, bytenr + num_bytes - 1, GFP_NOFS | __GFP_NOFAIL); - if (reserved) - trace_btrfs_reserved_extent_free(root, bytenr, num_bytes); return 0; } @@ -6476,6 +6675,9 @@ static int unpin_extent_range(struct btrfs_root *root, u64 start, u64 end, spin_lock(&cache->lock); cache->pinned -= len; space_info->bytes_pinned -= len; + + trace_btrfs_space_reservation(fs_info, "pinned", + space_info->flags, len, 0); space_info->max_extent_size = 0; percpu_counter_add(&space_info->total_bytes_pinned, -len); if (cache->ro) { @@ -6483,17 +6685,29 @@ static int unpin_extent_range(struct btrfs_root *root, u64 start, u64 end, readonly = true; } spin_unlock(&cache->lock); - if (!readonly && global_rsv->space_info == space_info) { + if (!readonly && return_free_space && + global_rsv->space_info == space_info) { + u64 to_add = len; + WARN_ON(!return_free_space); spin_lock(&global_rsv->lock); if (!global_rsv->full) { - len = min(len, global_rsv->size - - global_rsv->reserved); - global_rsv->reserved += len; - space_info->bytes_may_use += len; + to_add = min(len, global_rsv->size - + global_rsv->reserved); + global_rsv->reserved += to_add; + space_info->bytes_may_use += to_add; if (global_rsv->reserved >= global_rsv->size) global_rsv->full = 1; + trace_btrfs_space_reservation(fs_info, + "space_info", + space_info->flags, + to_add, 1); + len -= to_add; } spin_unlock(&global_rsv->lock); + /* Add to any tickets we may have */ + if (len) + space_info_add_new_bytes(fs_info, space_info, + len); } spin_unlock(&space_info->lock); } @@ -7782,12 +7996,10 @@ static int __btrfs_free_reserved_extent(struct btrfs_root *root, ret = btrfs_discard_extent(root, start, len, NULL); btrfs_add_free_space(cache, start, len); btrfs_update_reserved_bytes(cache, len, RESERVE_FREE, delalloc); + trace_btrfs_reserved_extent_free(root, start, len); } btrfs_put_block_group(cache); - - trace_btrfs_reserved_extent_free(root, start, len); - return ret; } @@ -9791,13 +10003,15 @@ int btrfs_free_block_groups(struct btrfs_fs_info *info) space_info = list_entry(info->space_info.next, struct btrfs_space_info, list); - if (btrfs_test_opt(info->tree_root, ENOSPC_DEBUG)) { - if (WARN_ON(space_info->bytes_pinned > 0 || + + /* + * Do not hide this behind enospc_debug, this is actually + * important and indicates a real bug if this happens. + */ + if (WARN_ON(space_info->bytes_pinned > 0 || space_info->bytes_reserved > 0 || - space_info->bytes_may_use > 0)) { - dump_space_info(space_info, 0, 0); - } - } + space_info->bytes_may_use > 0)) + dump_space_info(space_info, 0, 0); list_del(&space_info->list); for (i = 0; i < BTRFS_NR_RAID_TYPES; i++) { struct kobject *kobj; @@ -10005,9 +10219,10 @@ int btrfs_read_block_groups(struct btrfs_root *root) goto error; } + trace_btrfs_add_block_group(root->fs_info, cache, 0); ret = update_space_info(info, cache->flags, found_key.offset, btrfs_block_group_used(&cache->item), - &space_info); + cache->bytes_super, &space_info); if (ret) { btrfs_remove_free_space_cache(cache); spin_lock(&info->block_group_cache_lock); @@ -10020,9 +10235,6 @@ int btrfs_read_block_groups(struct btrfs_root *root) } cache->space_info = space_info; - spin_lock(&cache->space_info->lock); - cache->space_info->bytes_readonly += cache->bytes_super; - spin_unlock(&cache->space_info->lock); __link_block_group(space_info, cache); @@ -10114,7 +10326,6 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans, int ret; struct btrfs_root *extent_root; struct btrfs_block_group_cache *cache; - extent_root = root->fs_info->extent_root; btrfs_set_log_full_commit(root->fs_info, trans); @@ -10160,7 +10371,7 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans, * assigned to our block group, but don't update its counters just yet. * We want our bg to be added to the rbtree with its ->space_info set. */ - ret = update_space_info(root->fs_info, cache->flags, 0, 0, + ret = update_space_info(root->fs_info, cache->flags, 0, 0, 0, &cache->space_info); if (ret) { btrfs_remove_free_space_cache(cache); @@ -10179,8 +10390,9 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans, * Now that our block group has its ->space_info set and is inserted in * the rbtree, update the space info's counters. */ + trace_btrfs_add_block_group(root->fs_info, cache, 1); ret = update_space_info(root->fs_info, cache->flags, size, bytes_used, - &cache->space_info); + cache->bytes_super, &cache->space_info); if (ret) { btrfs_remove_free_space_cache(cache); spin_lock(&root->fs_info->block_group_cache_lock); @@ -10193,16 +10405,11 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans, } update_global_block_rsv(root->fs_info); - spin_lock(&cache->space_info->lock); - cache->space_info->bytes_readonly += cache->bytes_super; - spin_unlock(&cache->space_info->lock); - __link_block_group(cache->space_info, cache); list_add_tail(&cache->bg_list, &trans->new_bgs); set_avail_alloc_bits(extent_root->fs_info, type); - return 0; } @@ -10747,21 +10954,21 @@ int btrfs_init_space_info(struct btrfs_fs_info *fs_info) mixed = 1; flags = BTRFS_BLOCK_GROUP_SYSTEM; - ret = update_space_info(fs_info, flags, 0, 0, &space_info); + ret = update_space_info(fs_info, flags, 0, 0, 0, &space_info); if (ret) goto out; if (mixed) { flags = BTRFS_BLOCK_GROUP_METADATA | BTRFS_BLOCK_GROUP_DATA; - ret = update_space_info(fs_info, flags, 0, 0, &space_info); + ret = update_space_info(fs_info, flags, 0, 0, 0, &space_info); } else { flags = BTRFS_BLOCK_GROUP_METADATA; - ret = update_space_info(fs_info, flags, 0, 0, &space_info); + ret = update_space_info(fs_info, flags, 0, 0, 0, &space_info); if (ret) goto out; flags = BTRFS_BLOCK_GROUP_DATA; - ret = update_space_info(fs_info, flags, 0, 0, &space_info); + ret = update_space_info(fs_info, flags, 0, 0, 0, &space_info); } out: return ret; diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c index 2234e88cf674..bcfb4a27ddd4 100644 --- a/fs/btrfs/file.c +++ b/fs/btrfs/file.c @@ -1629,13 +1629,11 @@ again: * managed to copy. */ if (num_sectors > dirty_sectors) { - /* - * we round down because we don't want to count - * any partial blocks actually sent through the - * IO machines - */ - release_bytes = round_down(release_bytes - copied, - root->sectorsize); + + /* release everything except the sectors we dirtied */ + release_bytes -= dirty_sectors << + root->fs_info->sb->s_blocksize_bits; + if (copied > 0) { spin_lock(&BTRFS_I(inode)->lock); BTRFS_I(inode)->outstanding_extents++; @@ -2479,7 +2477,7 @@ static int btrfs_punch_hole(struct inode *inode, loff_t offset, loff_t len) } ret = btrfs_block_rsv_migrate(&root->fs_info->trans_block_rsv, rsv, - min_size); + min_size, 0); BUG_ON(ret); trans->block_rsv = rsv; @@ -2522,7 +2520,7 @@ static int btrfs_punch_hole(struct inode *inode, loff_t offset, loff_t len) } ret = btrfs_block_rsv_migrate(&root->fs_info->trans_block_rsv, - rsv, min_size); + rsv, min_size, 0); BUG_ON(ret); /* shouldn't happen */ trans->block_rsv = rsv; diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index df731c0ebec7..8078077d1090 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -5263,7 +5263,7 @@ void btrfs_evict_inode(struct inode *inode) if (steal_from_global) { if (!btrfs_check_space_for_delayed_refs(trans, root)) ret = btrfs_block_rsv_migrate(global_rsv, rsv, - min_size); + min_size, 0); else ret = -ENOSPC; } @@ -9116,7 +9116,7 @@ static int btrfs_truncate(struct inode *inode) /* Migrate the slack space for the truncate to our reserve */ ret = btrfs_block_rsv_migrate(&root->fs_info->trans_block_rsv, rsv, - min_size); + min_size, 0); BUG_ON(ret); /* @@ -9156,7 +9156,7 @@ static int btrfs_truncate(struct inode *inode) } ret = btrfs_block_rsv_migrate(&root->fs_info->trans_block_rsv, - rsv, min_size); + rsv, min_size, 0); BUG_ON(ret); /* shouldn't happen */ trans->block_rsv = rsv; } @@ -9177,7 +9177,6 @@ static int btrfs_truncate(struct inode *inode) ret = btrfs_end_transaction(trans, root); btrfs_btree_balance_dirty(root); } - out: btrfs_free_block_rsv(root, rsv); diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c index 0477dca154ed..fc067b07e31f 100644 --- a/fs/btrfs/relocation.c +++ b/fs/btrfs/relocation.c @@ -2604,25 +2604,28 @@ static int reserve_metadata_space(struct btrfs_trans_handle *trans, trans->block_rsv = rc->block_rsv; rc->reserved_bytes += num_bytes; + + /* + * We are under a transaction here so we can only do limited flushing. + * If we get an enospc just kick back -EAGAIN so we know to drop the + * transaction and try to refill when we can flush all the things. + */ ret = btrfs_block_rsv_refill(root, rc->block_rsv, num_bytes, - BTRFS_RESERVE_FLUSH_ALL); + BTRFS_RESERVE_FLUSH_LIMIT); if (ret) { - if (ret == -EAGAIN) { - tmp = rc->extent_root->nodesize * - RELOCATION_RESERVED_NODES; - while (tmp <= rc->reserved_bytes) - tmp <<= 1; - /* - * only one thread can access block_rsv at this point, - * so we don't need hold lock to protect block_rsv. - * we expand more reservation size here to allow enough - * space for relocation and we will return earlier in - * enospc case. - */ - rc->block_rsv->size = tmp + rc->extent_root->nodesize * - RELOCATION_RESERVED_NODES; - } - return ret; + tmp = rc->extent_root->nodesize * RELOCATION_RESERVED_NODES; + while (tmp <= rc->reserved_bytes) + tmp <<= 1; + /* + * only one thread can access block_rsv at this point, + * so we don't need hold lock to protect block_rsv. + * we expand more reservation size here to allow enough + * space for relocation and we will return eailer in + * enospc case. + */ + rc->block_rsv->size = tmp + rc->extent_root->nodesize * + RELOCATION_RESERVED_NODES; + return -EAGAIN; } return 0; @@ -3871,6 +3874,7 @@ static noinline_for_stack int prepare_to_relocate(struct reloc_control *rc) { struct btrfs_trans_handle *trans; + int ret; rc->block_rsv = btrfs_alloc_block_rsv(rc->extent_root, BTRFS_BLOCK_RSV_TEMP); @@ -3885,6 +3889,11 @@ int prepare_to_relocate(struct reloc_control *rc) rc->reserved_bytes = 0; rc->block_rsv->size = rc->extent_root->nodesize * RELOCATION_RESERVED_NODES; + ret = btrfs_block_rsv_refill(rc->extent_root, + rc->block_rsv, rc->block_rsv->size, + BTRFS_RESERVE_FLUSH_ALL); + if (ret) + return ret; rc->create_reloc_tree = 1; set_reloc_control(rc); @@ -4643,7 +4652,7 @@ int btrfs_reloc_post_snapshot(struct btrfs_trans_handle *trans, if (rc->merge_reloc_tree) { ret = btrfs_block_rsv_migrate(&pending->block_rsv, rc->block_rsv, - rc->nodes_relocated); + rc->nodes_relocated, 1); if (ret) return ret; } diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c index 26a9d10d75e9..d5b6f959a3c3 100644 --- a/fs/ceph/addr.c +++ b/fs/ceph/addr.c @@ -1730,7 +1730,8 @@ enum { POOL_WRITE = 2, }; -static int __ceph_pool_perm_get(struct ceph_inode_info *ci, u32 pool) +static int __ceph_pool_perm_get(struct ceph_inode_info *ci, + s64 pool, struct ceph_string *pool_ns) { struct ceph_fs_client *fsc = ceph_inode_to_client(&ci->vfs_inode); struct ceph_mds_client *mdsc = fsc->mdsc; @@ -1738,6 +1739,7 @@ static int __ceph_pool_perm_get(struct ceph_inode_info *ci, u32 pool) struct rb_node **p, *parent; struct ceph_pool_perm *perm; struct page **pages; + size_t pool_ns_len; int err = 0, err2 = 0, have = 0; down_read(&mdsc->pool_perm_rwsem); @@ -1749,17 +1751,31 @@ static int __ceph_pool_perm_get(struct ceph_inode_info *ci, u32 pool) else if (pool > perm->pool) p = &(*p)->rb_right; else { - have = perm->perm; - break; + int ret = ceph_compare_string(pool_ns, + perm->pool_ns, + perm->pool_ns_len); + if (ret < 0) + p = &(*p)->rb_left; + else if (ret > 0) + p = &(*p)->rb_right; + else { + have = perm->perm; + break; + } } } up_read(&mdsc->pool_perm_rwsem); if (*p) goto out; - dout("__ceph_pool_perm_get pool %u no perm cached\n", pool); + if (pool_ns) + dout("__ceph_pool_perm_get pool %lld ns %.*s no perm cached\n", + pool, (int)pool_ns->len, pool_ns->str); + else + dout("__ceph_pool_perm_get pool %lld no perm cached\n", pool); down_write(&mdsc->pool_perm_rwsem); + p = &mdsc->pool_perm_tree.rb_node; parent = NULL; while (*p) { parent = *p; @@ -1769,8 +1785,17 @@ static int __ceph_pool_perm_get(struct ceph_inode_info *ci, u32 pool) else if (pool > perm->pool) p = &(*p)->rb_right; else { - have = perm->perm; - break; + int ret = ceph_compare_string(pool_ns, + perm->pool_ns, + perm->pool_ns_len); + if (ret < 0) + p = &(*p)->rb_left; + else if (ret > 0) + p = &(*p)->rb_right; + else { + have = perm->perm; + break; + } } } if (*p) { @@ -1788,6 +1813,8 @@ static int __ceph_pool_perm_get(struct ceph_inode_info *ci, u32 pool) rd_req->r_flags = CEPH_OSD_FLAG_READ; osd_req_op_init(rd_req, 0, CEPH_OSD_OP_STAT, 0); rd_req->r_base_oloc.pool = pool; + if (pool_ns) + rd_req->r_base_oloc.pool_ns = ceph_get_string(pool_ns); ceph_oid_printf(&rd_req->r_base_oid, "%llx.00000000", ci->i_vino.ino); err = ceph_osdc_alloc_messages(rd_req, GFP_NOFS); @@ -1841,7 +1868,8 @@ static int __ceph_pool_perm_get(struct ceph_inode_info *ci, u32 pool) goto out_unlock; } - perm = kmalloc(sizeof(*perm), GFP_NOFS); + pool_ns_len = pool_ns ? pool_ns->len : 0; + perm = kmalloc(sizeof(*perm) + pool_ns_len + 1, GFP_NOFS); if (!perm) { err = -ENOMEM; goto out_unlock; @@ -1849,6 +1877,11 @@ static int __ceph_pool_perm_get(struct ceph_inode_info *ci, u32 pool) perm->pool = pool; perm->perm = have; + perm->pool_ns_len = pool_ns_len; + if (pool_ns_len > 0) + memcpy(perm->pool_ns, pool_ns->str, pool_ns_len); + perm->pool_ns[pool_ns_len] = 0; + rb_link_node(&perm->node, parent, p); rb_insert_color(&perm->node, &mdsc->pool_perm_tree); err = 0; @@ -1860,43 +1893,46 @@ out_unlock: out: if (!err) err = have; - dout("__ceph_pool_perm_get pool %u result = %d\n", pool, err); + if (pool_ns) + dout("__ceph_pool_perm_get pool %lld ns %.*s result = %d\n", + pool, (int)pool_ns->len, pool_ns->str, err); + else + dout("__ceph_pool_perm_get pool %lld result = %d\n", pool, err); return err; } int ceph_pool_perm_check(struct ceph_inode_info *ci, int need) { - u32 pool; + s64 pool; + struct ceph_string *pool_ns; int ret, flags; - /* does not support pool namespace yet */ - if (ci->i_pool_ns_len) - return -EIO; - if (ceph_test_mount_opt(ceph_inode_to_client(&ci->vfs_inode), NOPOOLPERM)) return 0; spin_lock(&ci->i_ceph_lock); flags = ci->i_ceph_flags; - pool = ceph_file_layout_pg_pool(ci->i_layout); + pool = ci->i_layout.pool_id; spin_unlock(&ci->i_ceph_lock); check: if (flags & CEPH_I_POOL_PERM) { if ((need & CEPH_CAP_FILE_RD) && !(flags & CEPH_I_POOL_RD)) { - dout("ceph_pool_perm_check pool %u no read perm\n", + dout("ceph_pool_perm_check pool %lld no read perm\n", pool); return -EPERM; } if ((need & CEPH_CAP_FILE_WR) && !(flags & CEPH_I_POOL_WR)) { - dout("ceph_pool_perm_check pool %u no write perm\n", + dout("ceph_pool_perm_check pool %lld no write perm\n", pool); return -EPERM; } return 0; } - ret = __ceph_pool_perm_get(ci, pool); + pool_ns = ceph_try_get_string(ci->i_layout.pool_ns); + ret = __ceph_pool_perm_get(ci, pool, pool_ns); + ceph_put_string(pool_ns); if (ret < 0) return ret; @@ -1907,10 +1943,11 @@ check: flags |= CEPH_I_POOL_WR; spin_lock(&ci->i_ceph_lock); - if (pool == ceph_file_layout_pg_pool(ci->i_layout)) { - ci->i_ceph_flags = flags; + if (pool == ci->i_layout.pool_id && + pool_ns == rcu_dereference_raw(ci->i_layout.pool_ns)) { + ci->i_ceph_flags |= flags; } else { - pool = ceph_file_layout_pg_pool(ci->i_layout); + pool = ci->i_layout.pool_id; flags = ci->i_ceph_flags; } spin_unlock(&ci->i_ceph_lock); diff --git a/fs/ceph/cache.c b/fs/ceph/cache.c index 238c55b01723..5bc5d37b1217 100644 --- a/fs/ceph/cache.c +++ b/fs/ceph/cache.c @@ -71,7 +71,7 @@ int ceph_fscache_register_fs(struct ceph_fs_client* fsc) &ceph_fscache_fsid_object_def, fsc, true); if (!fsc->fscache) - pr_err("Unable to resgister fsid: %p fscache cookie", fsc); + pr_err("Unable to register fsid: %p fscache cookie\n", fsc); return 0; } diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c index 6f60d0a3d0f9..99115cae1652 100644 --- a/fs/ceph/caps.c +++ b/fs/ceph/caps.c @@ -40,6 +40,11 @@ * cluster to release server state. */ +static u64 __get_oldest_flush_tid(struct ceph_mds_client *mdsc); +static void __kick_flushing_caps(struct ceph_mds_client *mdsc, + struct ceph_mds_session *session, + struct ceph_inode_info *ci, + u64 oldest_flush_tid); /* * Generate readable cap strings for debugging output. @@ -849,12 +854,14 @@ int __ceph_caps_used(struct ceph_inode_info *ci) */ int __ceph_caps_file_wanted(struct ceph_inode_info *ci) { - int want = 0; - int mode; - for (mode = 0; mode < CEPH_FILE_MODE_NUM; mode++) - if (ci->i_nr_by_mode[mode]) - want |= ceph_caps_for_mode(mode); - return want; + int i, bits = 0; + for (i = 0; i < CEPH_FILE_MODE_BITS; i++) { + if (ci->i_nr_by_mode[i]) + bits |= 1 << i; + } + if (bits == 0) + return 0; + return ceph_caps_for_mode(bits >> 1); } /* @@ -991,7 +998,7 @@ static int send_cap_msg(struct ceph_mds_session *session, u32 seq, u64 flush_tid, u64 oldest_flush_tid, u32 issue_seq, u32 mseq, u64 size, u64 max_size, struct timespec *mtime, struct timespec *atime, - struct timespec *ctime, u64 time_warp_seq, + struct timespec *ctime, u32 time_warp_seq, kuid_t uid, kgid_t gid, umode_t mode, u64 xattr_version, struct ceph_buffer *xattrs_buf, @@ -1116,8 +1123,8 @@ static int __send_cap(struct ceph_mds_client *mdsc, struct ceph_cap *cap, struct inode *inode = &ci->vfs_inode; u64 cap_id = cap->cap_id; int held, revoking, dropping, keep; - u64 seq, issue_seq, mseq, time_warp_seq, follows; - u64 size, max_size; + u64 follows, size, max_size; + u32 seq, issue_seq, mseq, time_warp_seq; struct timespec mtime, atime, ctime; int wake = 0; umode_t mode; @@ -1215,6 +1222,22 @@ static int __send_cap(struct ceph_mds_client *mdsc, struct ceph_cap *cap, return delayed; } +static inline int __send_flush_snap(struct inode *inode, + struct ceph_mds_session *session, + struct ceph_cap_snap *capsnap, + u32 mseq, u64 oldest_flush_tid) +{ + return send_cap_msg(session, ceph_vino(inode).ino, 0, + CEPH_CAP_OP_FLUSHSNAP, capsnap->issued, 0, + capsnap->dirty, 0, capsnap->cap_flush.tid, + oldest_flush_tid, 0, mseq, capsnap->size, 0, + &capsnap->mtime, &capsnap->atime, + &capsnap->ctime, capsnap->time_warp_seq, + capsnap->uid, capsnap->gid, capsnap->mode, + capsnap->xattr_version, capsnap->xattr_blob, + capsnap->follows, capsnap->inline_data); +} + /* * When a snapshot is taken, clients accumulate dirty metadata on * inodes with capabilities in ceph_cap_snaps to describe the file @@ -1222,37 +1245,22 @@ static int __send_cap(struct ceph_mds_client *mdsc, struct ceph_cap *cap, * asynchronously back to the MDS once sync writes complete and dirty * data is written out. * - * Unless @kick is true, skip cap_snaps that were already sent to - * the MDS (i.e., during this session). - * * Called under i_ceph_lock. Takes s_mutex as needed. */ -void __ceph_flush_snaps(struct ceph_inode_info *ci, - struct ceph_mds_session **psession, - int kick) +static void __ceph_flush_snaps(struct ceph_inode_info *ci, + struct ceph_mds_session *session) __releases(ci->i_ceph_lock) __acquires(ci->i_ceph_lock) { struct inode *inode = &ci->vfs_inode; - int mds; + struct ceph_mds_client *mdsc = session->s_mdsc; struct ceph_cap_snap *capsnap; - u32 mseq; - struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc; - struct ceph_mds_session *session = NULL; /* if session != NULL, we hold - session->s_mutex */ - u64 next_follows = 0; /* keep track of how far we've gotten through the - i_cap_snaps list, and skip these entries next time - around to avoid an infinite loop */ + u64 oldest_flush_tid = 0; + u64 first_tid = 1, last_tid = 0; - if (psession) - session = *psession; + dout("__flush_snaps %p session %p\n", inode, session); - dout("__flush_snaps %p\n", inode); -retry: list_for_each_entry(capsnap, &ci->i_cap_snaps, ci_item) { - /* avoid an infiniute loop after retry */ - if (capsnap->follows < next_follows) - continue; /* * we need to wait for sync writes to complete and for dirty * pages to be written out. @@ -1263,97 +1271,129 @@ retry: /* should be removed by ceph_try_drop_cap_snap() */ BUG_ON(!capsnap->need_flush); - /* pick mds, take s_mutex */ - if (ci->i_auth_cap == NULL) { - dout("no auth cap (migrating?), doing nothing\n"); - goto out; - } - /* only flush each capsnap once */ - if (!kick && !list_empty(&capsnap->flushing_item)) { - dout("already flushed %p, skipping\n", capsnap); + if (capsnap->cap_flush.tid > 0) { + dout(" already flushed %p, skipping\n", capsnap); continue; } - mds = ci->i_auth_cap->session->s_mds; - mseq = ci->i_auth_cap->mseq; + spin_lock(&mdsc->cap_dirty_lock); + capsnap->cap_flush.tid = ++mdsc->last_cap_flush_tid; + list_add_tail(&capsnap->cap_flush.g_list, + &mdsc->cap_flush_list); + if (oldest_flush_tid == 0) + oldest_flush_tid = __get_oldest_flush_tid(mdsc); + if (list_empty(&ci->i_flushing_item)) { + list_add_tail(&ci->i_flushing_item, + &session->s_cap_flushing); + } + spin_unlock(&mdsc->cap_dirty_lock); + + list_add_tail(&capsnap->cap_flush.i_list, + &ci->i_cap_flush_list); - if (session && session->s_mds != mds) { - dout("oops, wrong session %p mutex\n", session); - if (kick) - goto out; + if (first_tid == 1) + first_tid = capsnap->cap_flush.tid; + last_tid = capsnap->cap_flush.tid; + } - mutex_unlock(&session->s_mutex); - ceph_put_mds_session(session); - session = NULL; + ci->i_ceph_flags &= ~CEPH_I_FLUSH_SNAPS; + + while (first_tid <= last_tid) { + struct ceph_cap *cap = ci->i_auth_cap; + struct ceph_cap_flush *cf; + int ret; + + if (!(cap && cap->session == session)) { + dout("__flush_snaps %p auth cap %p not mds%d, " + "stop\n", inode, cap, session->s_mds); + break; } - if (!session) { - spin_unlock(&ci->i_ceph_lock); - mutex_lock(&mdsc->mutex); - session = __ceph_lookup_mds_session(mdsc, mds); - mutex_unlock(&mdsc->mutex); - if (session) { - dout("inverting session/ino locks on %p\n", - session); - mutex_lock(&session->s_mutex); + + ret = -ENOENT; + list_for_each_entry(cf, &ci->i_cap_flush_list, i_list) { + if (cf->tid >= first_tid) { + ret = 0; + break; } - /* - * if session == NULL, we raced against a cap - * deletion or migration. retry, and we'll - * get a better @mds value next time. - */ - spin_lock(&ci->i_ceph_lock); - goto retry; } + if (ret < 0) + break; - spin_lock(&mdsc->cap_dirty_lock); - capsnap->flush_tid = ++mdsc->last_cap_flush_tid; - spin_unlock(&mdsc->cap_dirty_lock); + first_tid = cf->tid + 1; + capsnap = container_of(cf, struct ceph_cap_snap, cap_flush); atomic_inc(&capsnap->nref); - if (list_empty(&capsnap->flushing_item)) - list_add_tail(&capsnap->flushing_item, - &session->s_cap_snaps_flushing); spin_unlock(&ci->i_ceph_lock); - dout("flush_snaps %p cap_snap %p follows %lld tid %llu\n", - inode, capsnap, capsnap->follows, capsnap->flush_tid); - send_cap_msg(session, ceph_vino(inode).ino, 0, - CEPH_CAP_OP_FLUSHSNAP, capsnap->issued, 0, - capsnap->dirty, 0, capsnap->flush_tid, 0, - 0, mseq, capsnap->size, 0, - &capsnap->mtime, &capsnap->atime, - &capsnap->ctime, capsnap->time_warp_seq, - capsnap->uid, capsnap->gid, capsnap->mode, - capsnap->xattr_version, capsnap->xattr_blob, - capsnap->follows, capsnap->inline_data); - - next_follows = capsnap->follows + 1; - ceph_put_cap_snap(capsnap); + dout("__flush_snaps %p capsnap %p tid %llu %s\n", + inode, capsnap, cf->tid, ceph_cap_string(capsnap->dirty)); + + ret = __send_flush_snap(inode, session, capsnap, cap->mseq, + oldest_flush_tid); + if (ret < 0) { + pr_err("__flush_snaps: error sending cap flushsnap, " + "ino (%llx.%llx) tid %llu follows %llu\n", + ceph_vinop(inode), cf->tid, capsnap->follows); + } + ceph_put_cap_snap(capsnap); spin_lock(&ci->i_ceph_lock); - goto retry; } +} - /* we flushed them all; remove this inode from the queue */ - spin_lock(&mdsc->snap_flush_lock); - list_del_init(&ci->i_snap_flush_item); - spin_unlock(&mdsc->snap_flush_lock); +void ceph_flush_snaps(struct ceph_inode_info *ci, + struct ceph_mds_session **psession) +{ + struct inode *inode = &ci->vfs_inode; + struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc; + struct ceph_mds_session *session = *psession; + int mds; + dout("ceph_flush_snaps %p\n", inode); +retry: + spin_lock(&ci->i_ceph_lock); + if (!(ci->i_ceph_flags & CEPH_I_FLUSH_SNAPS)) { + dout(" no capsnap needs flush, doing nothing\n"); + goto out; + } + if (!ci->i_auth_cap) { + dout(" no auth cap (migrating?), doing nothing\n"); + goto out; + } -out: - if (psession) - *psession = session; - else if (session) { + mds = ci->i_auth_cap->session->s_mds; + if (session && session->s_mds != mds) { + dout(" oops, wrong session %p mutex\n", session); mutex_unlock(&session->s_mutex); ceph_put_mds_session(session); + session = NULL; + } + if (!session) { + spin_unlock(&ci->i_ceph_lock); + mutex_lock(&mdsc->mutex); + session = __ceph_lookup_mds_session(mdsc, mds); + mutex_unlock(&mdsc->mutex); + if (session) { + dout(" inverting session/ino locks on %p\n", session); + mutex_lock(&session->s_mutex); + } + goto retry; } -} -static void ceph_flush_snaps(struct ceph_inode_info *ci) -{ - spin_lock(&ci->i_ceph_lock); - __ceph_flush_snaps(ci, NULL, 0); + __ceph_flush_snaps(ci, session); +out: spin_unlock(&ci->i_ceph_lock); + + if (psession) { + *psession = session; + } else { + mutex_unlock(&session->s_mutex); + ceph_put_mds_session(session); + } + /* we flushed them all; remove this inode from the queue */ + spin_lock(&mdsc->snap_flush_lock); + list_del_init(&ci->i_snap_flush_item); + spin_unlock(&mdsc->snap_flush_lock); } /* @@ -1411,52 +1451,6 @@ int __ceph_mark_dirty_caps(struct ceph_inode_info *ci, int mask, return dirty; } -static void __add_cap_flushing_to_inode(struct ceph_inode_info *ci, - struct ceph_cap_flush *cf) -{ - struct rb_node **p = &ci->i_cap_flush_tree.rb_node; - struct rb_node *parent = NULL; - struct ceph_cap_flush *other = NULL; - - while (*p) { - parent = *p; - other = rb_entry(parent, struct ceph_cap_flush, i_node); - - if (cf->tid < other->tid) - p = &(*p)->rb_left; - else if (cf->tid > other->tid) - p = &(*p)->rb_right; - else - BUG(); - } - - rb_link_node(&cf->i_node, parent, p); - rb_insert_color(&cf->i_node, &ci->i_cap_flush_tree); -} - -static void __add_cap_flushing_to_mdsc(struct ceph_mds_client *mdsc, - struct ceph_cap_flush *cf) -{ - struct rb_node **p = &mdsc->cap_flush_tree.rb_node; - struct rb_node *parent = NULL; - struct ceph_cap_flush *other = NULL; - - while (*p) { - parent = *p; - other = rb_entry(parent, struct ceph_cap_flush, g_node); - - if (cf->tid < other->tid) - p = &(*p)->rb_left; - else if (cf->tid > other->tid) - p = &(*p)->rb_right; - else - BUG(); - } - - rb_link_node(&cf->g_node, parent, p); - rb_insert_color(&cf->g_node, &mdsc->cap_flush_tree); -} - struct ceph_cap_flush *ceph_alloc_cap_flush(void) { return kmem_cache_alloc(ceph_cap_flush_cachep, GFP_KERNEL); @@ -1470,23 +1464,54 @@ void ceph_free_cap_flush(struct ceph_cap_flush *cf) static u64 __get_oldest_flush_tid(struct ceph_mds_client *mdsc) { - struct rb_node *n = rb_first(&mdsc->cap_flush_tree); - if (n) { + if (!list_empty(&mdsc->cap_flush_list)) { struct ceph_cap_flush *cf = - rb_entry(n, struct ceph_cap_flush, g_node); + list_first_entry(&mdsc->cap_flush_list, + struct ceph_cap_flush, g_list); return cf->tid; } return 0; } /* + * Remove cap_flush from the mdsc's or inode's flushing cap list. + * Return true if caller needs to wake up flush waiters. + */ +static bool __finish_cap_flush(struct ceph_mds_client *mdsc, + struct ceph_inode_info *ci, + struct ceph_cap_flush *cf) +{ + struct ceph_cap_flush *prev; + bool wake = cf->wake; + if (mdsc) { + /* are there older pending cap flushes? */ + if (wake && cf->g_list.prev != &mdsc->cap_flush_list) { + prev = list_prev_entry(cf, g_list); + prev->wake = true; + wake = false; + } + list_del(&cf->g_list); + } else if (ci) { + if (wake && cf->i_list.prev != &ci->i_cap_flush_list) { + prev = list_prev_entry(cf, i_list); + prev->wake = true; + wake = false; + } + list_del(&cf->i_list); + } else { + BUG_ON(1); + } + return wake; +} + +/* * Add dirty inode to the flushing list. Assigned a seq number so we * can wait for caps to flush without starving. * * Called under i_ceph_lock. */ static int __mark_caps_flushing(struct inode *inode, - struct ceph_mds_session *session, + struct ceph_mds_session *session, bool wake, u64 *flush_tid, u64 *oldest_flush_tid) { struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc; @@ -1509,26 +1534,22 @@ static int __mark_caps_flushing(struct inode *inode, swap(cf, ci->i_prealloc_cap_flush); cf->caps = flushing; + cf->wake = wake; spin_lock(&mdsc->cap_dirty_lock); list_del_init(&ci->i_dirty_item); cf->tid = ++mdsc->last_cap_flush_tid; - __add_cap_flushing_to_mdsc(mdsc, cf); + list_add_tail(&cf->g_list, &mdsc->cap_flush_list); *oldest_flush_tid = __get_oldest_flush_tid(mdsc); if (list_empty(&ci->i_flushing_item)) { list_add_tail(&ci->i_flushing_item, &session->s_cap_flushing); mdsc->num_cap_flushing++; - dout(" inode %p now flushing tid %llu\n", inode, cf->tid); - } else { - list_move_tail(&ci->i_flushing_item, &session->s_cap_flushing); - dout(" inode %p now flushing (more) tid %llu\n", - inode, cf->tid); } spin_unlock(&mdsc->cap_dirty_lock); - __add_cap_flushing_to_inode(ci, cf); + list_add_tail(&cf->i_list, &ci->i_cap_flush_list); *flush_tid = cf->tid; return flushing; @@ -1583,10 +1604,11 @@ void ceph_check_caps(struct ceph_inode_info *ci, int flags, int mds = -1; /* keep track of how far we've gone through i_caps list to avoid an infinite loop on retry */ struct rb_node *p; - int tried_invalidate = 0; - int delayed = 0, sent = 0, force_requeue = 0, num; - int queue_invalidate = 0; - int is_delayed = flags & CHECK_CAPS_NODELAY; + int delayed = 0, sent = 0, num; + bool is_delayed = flags & CHECK_CAPS_NODELAY; + bool queue_invalidate = false; + bool force_requeue = false; + bool tried_invalidate = false; /* if we are unmounting, flush any unused caps immediately. */ if (mdsc->stopping) @@ -1597,9 +1619,6 @@ void ceph_check_caps(struct ceph_inode_info *ci, int flags, if (ci->i_ceph_flags & CEPH_I_FLUSH) flags |= CHECK_CAPS_FLUSH; - /* flush snaps first time around only */ - if (!list_empty(&ci->i_cap_snaps)) - __ceph_flush_snaps(ci, &session, 0); goto retry_locked; retry: spin_lock(&ci->i_ceph_lock); @@ -1666,17 +1685,17 @@ retry_locked: if (revoking & (CEPH_CAP_FILE_CACHE| CEPH_CAP_FILE_LAZYIO)) { dout("check_caps queuing invalidate\n"); - queue_invalidate = 1; + queue_invalidate = true; ci->i_rdcache_revoking = ci->i_rdcache_gen; } else { dout("check_caps failed to invalidate pages\n"); /* we failed to invalidate pages. check these caps again later. */ - force_requeue = 1; + force_requeue = true; __cap_set_timeouts(mdsc, ci); } } - tried_invalidate = 1; + tried_invalidate = true; goto retry_locked; } @@ -1720,10 +1739,15 @@ retry_locked: } } /* flush anything dirty? */ - if (cap == ci->i_auth_cap && (flags & CHECK_CAPS_FLUSH) && - ci->i_dirty_caps) { - dout("flushing dirty caps\n"); - goto ack; + if (cap == ci->i_auth_cap) { + if ((flags & CHECK_CAPS_FLUSH) && ci->i_dirty_caps) { + dout("flushing dirty caps\n"); + goto ack; + } + if (ci->i_ceph_flags & CEPH_I_FLUSH_SNAPS) { + dout("flushing snap caps\n"); + goto ack; + } } /* completed revocation? going down and there are no caps? */ @@ -1782,6 +1806,26 @@ ack: goto retry; } } + + /* kick flushing and flush snaps before sending normal + * cap message */ + if (cap == ci->i_auth_cap && + (ci->i_ceph_flags & + (CEPH_I_KICK_FLUSH | CEPH_I_FLUSH_SNAPS))) { + if (ci->i_ceph_flags & CEPH_I_KICK_FLUSH) { + spin_lock(&mdsc->cap_dirty_lock); + oldest_flush_tid = __get_oldest_flush_tid(mdsc); + spin_unlock(&mdsc->cap_dirty_lock); + __kick_flushing_caps(mdsc, session, ci, + oldest_flush_tid); + ci->i_ceph_flags &= ~CEPH_I_KICK_FLUSH; + } + if (ci->i_ceph_flags & CEPH_I_FLUSH_SNAPS) + __ceph_flush_snaps(ci, session); + + goto retry_locked; + } + /* take snap_rwsem after session mutex */ if (!took_snap_rwsem) { if (down_read_trylock(&mdsc->snap_rwsem) == 0) { @@ -1796,7 +1840,7 @@ ack: } if (cap == ci->i_auth_cap && ci->i_dirty_caps) { - flushing = __mark_caps_flushing(inode, session, + flushing = __mark_caps_flushing(inode, session, false, &flush_tid, &oldest_flush_tid); } else { @@ -1822,7 +1866,7 @@ ack: * otherwise cancel. */ if (delayed && is_delayed) - force_requeue = 1; /* __send_cap delayed release; requeue */ + force_requeue = true; /* __send_cap delayed release; requeue */ if (!delayed && !is_delayed) __cap_delay_cancel(mdsc, ci); else if (!is_delayed || force_requeue) @@ -1873,8 +1917,8 @@ retry: if (cap->session->s_state < CEPH_MDS_SESSION_OPEN) goto out; - flushing = __mark_caps_flushing(inode, session, &flush_tid, - &oldest_flush_tid); + flushing = __mark_caps_flushing(inode, session, true, + &flush_tid, &oldest_flush_tid); /* __send_cap drops i_ceph_lock */ delayed = __send_cap(mdsc, cap, CEPH_CAP_OP_FLUSH, used, want, @@ -1887,10 +1931,11 @@ retry: spin_unlock(&ci->i_ceph_lock); } } else { - struct rb_node *n = rb_last(&ci->i_cap_flush_tree); - if (n) { + if (!list_empty(&ci->i_cap_flush_list)) { struct ceph_cap_flush *cf = - rb_entry(n, struct ceph_cap_flush, i_node); + list_last_entry(&ci->i_cap_flush_list, + struct ceph_cap_flush, i_list); + cf->wake = true; flush_tid = cf->tid; } flushing = ci->i_flushing_caps; @@ -1910,14 +1955,13 @@ out: static int caps_are_flushed(struct inode *inode, u64 flush_tid) { struct ceph_inode_info *ci = ceph_inode(inode); - struct ceph_cap_flush *cf; - struct rb_node *n; int ret = 1; spin_lock(&ci->i_ceph_lock); - n = rb_first(&ci->i_cap_flush_tree); - if (n) { - cf = rb_entry(n, struct ceph_cap_flush, i_node); + if (!list_empty(&ci->i_cap_flush_list)) { + struct ceph_cap_flush * cf = + list_first_entry(&ci->i_cap_flush_list, + struct ceph_cap_flush, i_list); if (cf->tid <= flush_tid) ret = 0; } @@ -1926,53 +1970,6 @@ static int caps_are_flushed(struct inode *inode, u64 flush_tid) } /* - * Wait on any unsafe replies for the given inode. First wait on the - * newest request, and make that the upper bound. Then, if there are - * more requests, keep waiting on the oldest as long as it is still older - * than the original request. - */ -static void sync_write_wait(struct inode *inode) -{ - struct ceph_inode_info *ci = ceph_inode(inode); - struct list_head *head = &ci->i_unsafe_writes; - struct ceph_osd_request *req; - u64 last_tid; - - if (!S_ISREG(inode->i_mode)) - return; - - spin_lock(&ci->i_unsafe_lock); - if (list_empty(head)) - goto out; - - /* set upper bound as _last_ entry in chain */ - req = list_last_entry(head, struct ceph_osd_request, - r_unsafe_item); - last_tid = req->r_tid; - - do { - ceph_osdc_get_request(req); - spin_unlock(&ci->i_unsafe_lock); - dout("sync_write_wait on tid %llu (until %llu)\n", - req->r_tid, last_tid); - wait_for_completion(&req->r_safe_completion); - spin_lock(&ci->i_unsafe_lock); - ceph_osdc_put_request(req); - - /* - * from here on look at first entry in chain, since we - * only want to wait for anything older than last_tid - */ - if (list_empty(head)) - break; - req = list_first_entry(head, struct ceph_osd_request, - r_unsafe_item); - } while (req->r_tid < last_tid); -out: - spin_unlock(&ci->i_unsafe_lock); -} - -/* * wait for any unsafe requests to complete. */ static int unsafe_request_wait(struct inode *inode) @@ -2024,7 +2021,8 @@ int ceph_fsync(struct file *file, loff_t start, loff_t end, int datasync) int dirty; dout("fsync %p%s\n", inode, datasync ? " datasync" : ""); - sync_write_wait(inode); + + ceph_sync_write_wait(inode); ret = filemap_write_and_wait_range(inode->i_mapping, start, end); if (ret < 0) @@ -2087,87 +2085,74 @@ int ceph_write_inode(struct inode *inode, struct writeback_control *wbc) return err; } -/* - * After a recovering MDS goes active, we need to resend any caps - * we were flushing. - * - * Caller holds session->s_mutex. - */ -static void kick_flushing_capsnaps(struct ceph_mds_client *mdsc, - struct ceph_mds_session *session) -{ - struct ceph_cap_snap *capsnap; - - dout("kick_flushing_capsnaps mds%d\n", session->s_mds); - list_for_each_entry(capsnap, &session->s_cap_snaps_flushing, - flushing_item) { - struct ceph_inode_info *ci = capsnap->ci; - struct inode *inode = &ci->vfs_inode; - struct ceph_cap *cap; - - spin_lock(&ci->i_ceph_lock); - cap = ci->i_auth_cap; - if (cap && cap->session == session) { - dout("kick_flushing_caps %p cap %p capsnap %p\n", inode, - cap, capsnap); - __ceph_flush_snaps(ci, &session, 1); - } else { - pr_err("%p auth cap %p not mds%d ???\n", inode, - cap, session->s_mds); - } - spin_unlock(&ci->i_ceph_lock); - } -} - -static int __kick_flushing_caps(struct ceph_mds_client *mdsc, - struct ceph_mds_session *session, - struct ceph_inode_info *ci) +static void __kick_flushing_caps(struct ceph_mds_client *mdsc, + struct ceph_mds_session *session, + struct ceph_inode_info *ci, + u64 oldest_flush_tid) + __releases(ci->i_ceph_lock) + __acquires(ci->i_ceph_lock) { struct inode *inode = &ci->vfs_inode; struct ceph_cap *cap; struct ceph_cap_flush *cf; - struct rb_node *n; - int delayed = 0; + int ret; u64 first_tid = 0; - u64 oldest_flush_tid; - spin_lock(&mdsc->cap_dirty_lock); - oldest_flush_tid = __get_oldest_flush_tid(mdsc); - spin_unlock(&mdsc->cap_dirty_lock); + list_for_each_entry(cf, &ci->i_cap_flush_list, i_list) { + if (cf->tid < first_tid) + continue; - while (true) { - spin_lock(&ci->i_ceph_lock); cap = ci->i_auth_cap; if (!(cap && cap->session == session)) { - pr_err("%p auth cap %p not mds%d ???\n", inode, - cap, session->s_mds); - spin_unlock(&ci->i_ceph_lock); + pr_err("%p auth cap %p not mds%d ???\n", + inode, cap, session->s_mds); break; } - for (n = rb_first(&ci->i_cap_flush_tree); n; n = rb_next(n)) { - cf = rb_entry(n, struct ceph_cap_flush, i_node); - if (cf->tid >= first_tid) - break; - } - if (!n) { + first_tid = cf->tid + 1; + + if (cf->caps) { + dout("kick_flushing_caps %p cap %p tid %llu %s\n", + inode, cap, cf->tid, ceph_cap_string(cf->caps)); + ci->i_ceph_flags |= CEPH_I_NODELAY; + ret = __send_cap(mdsc, cap, CEPH_CAP_OP_FLUSH, + __ceph_caps_used(ci), + __ceph_caps_wanted(ci), + cap->issued | cap->implemented, + cf->caps, cf->tid, oldest_flush_tid); + if (ret) { + pr_err("kick_flushing_caps: error sending " + "cap flush, ino (%llx.%llx) " + "tid %llu flushing %s\n", + ceph_vinop(inode), cf->tid, + ceph_cap_string(cf->caps)); + } + } else { + struct ceph_cap_snap *capsnap = + container_of(cf, struct ceph_cap_snap, + cap_flush); + dout("kick_flushing_caps %p capsnap %p tid %llu %s\n", + inode, capsnap, cf->tid, + ceph_cap_string(capsnap->dirty)); + + atomic_inc(&capsnap->nref); spin_unlock(&ci->i_ceph_lock); - break; - } - cf = rb_entry(n, struct ceph_cap_flush, i_node); + ret = __send_flush_snap(inode, session, capsnap, cap->mseq, + oldest_flush_tid); + if (ret < 0) { + pr_err("kick_flushing_caps: error sending " + "cap flushsnap, ino (%llx.%llx) " + "tid %llu follows %llu\n", + ceph_vinop(inode), cf->tid, + capsnap->follows); + } - first_tid = cf->tid + 1; + ceph_put_cap_snap(capsnap); + } - dout("kick_flushing_caps %p cap %p tid %llu %s\n", inode, - cap, cf->tid, ceph_cap_string(cf->caps)); - delayed |= __send_cap(mdsc, cap, CEPH_CAP_OP_FLUSH, - __ceph_caps_used(ci), - __ceph_caps_wanted(ci), - cap->issued | cap->implemented, - cf->caps, cf->tid, oldest_flush_tid); + spin_lock(&ci->i_ceph_lock); } - return delayed; } void ceph_early_kick_flushing_caps(struct ceph_mds_client *mdsc, @@ -2175,8 +2160,14 @@ void ceph_early_kick_flushing_caps(struct ceph_mds_client *mdsc, { struct ceph_inode_info *ci; struct ceph_cap *cap; + u64 oldest_flush_tid; dout("early_kick_flushing_caps mds%d\n", session->s_mds); + + spin_lock(&mdsc->cap_dirty_lock); + oldest_flush_tid = __get_oldest_flush_tid(mdsc); + spin_unlock(&mdsc->cap_dirty_lock); + list_for_each_entry(ci, &session->s_cap_flushing, i_flushing_item) { spin_lock(&ci->i_ceph_lock); cap = ci->i_auth_cap; @@ -2196,10 +2187,11 @@ void ceph_early_kick_flushing_caps(struct ceph_mds_client *mdsc, */ if ((cap->issued & ci->i_flushing_caps) != ci->i_flushing_caps) { - spin_unlock(&ci->i_ceph_lock); - if (!__kick_flushing_caps(mdsc, session, ci)) - continue; - spin_lock(&ci->i_ceph_lock); + ci->i_ceph_flags &= ~CEPH_I_KICK_FLUSH; + __kick_flushing_caps(mdsc, session, ci, + oldest_flush_tid); + } else { + ci->i_ceph_flags |= CEPH_I_KICK_FLUSH; } spin_unlock(&ci->i_ceph_lock); @@ -2210,50 +2202,56 @@ void ceph_kick_flushing_caps(struct ceph_mds_client *mdsc, struct ceph_mds_session *session) { struct ceph_inode_info *ci; - - kick_flushing_capsnaps(mdsc, session); + struct ceph_cap *cap; + u64 oldest_flush_tid; dout("kick_flushing_caps mds%d\n", session->s_mds); + + spin_lock(&mdsc->cap_dirty_lock); + oldest_flush_tid = __get_oldest_flush_tid(mdsc); + spin_unlock(&mdsc->cap_dirty_lock); + list_for_each_entry(ci, &session->s_cap_flushing, i_flushing_item) { - int delayed = __kick_flushing_caps(mdsc, session, ci); - if (delayed) { - spin_lock(&ci->i_ceph_lock); - __cap_delay_requeue(mdsc, ci); + spin_lock(&ci->i_ceph_lock); + cap = ci->i_auth_cap; + if (!(cap && cap->session == session)) { + pr_err("%p auth cap %p not mds%d ???\n", + &ci->vfs_inode, cap, session->s_mds); spin_unlock(&ci->i_ceph_lock); + continue; + } + if (ci->i_ceph_flags & CEPH_I_KICK_FLUSH) { + ci->i_ceph_flags &= ~CEPH_I_KICK_FLUSH; + __kick_flushing_caps(mdsc, session, ci, + oldest_flush_tid); } + spin_unlock(&ci->i_ceph_lock); } } static void kick_flushing_inode_caps(struct ceph_mds_client *mdsc, struct ceph_mds_session *session, struct inode *inode) + __releases(ci->i_ceph_lock) { struct ceph_inode_info *ci = ceph_inode(inode); struct ceph_cap *cap; - spin_lock(&ci->i_ceph_lock); cap = ci->i_auth_cap; dout("kick_flushing_inode_caps %p flushing %s\n", inode, ceph_cap_string(ci->i_flushing_caps)); - __ceph_flush_snaps(ci, &session, 1); - - if (ci->i_flushing_caps) { - int delayed; - + if (!list_empty(&ci->i_cap_flush_list)) { + u64 oldest_flush_tid; spin_lock(&mdsc->cap_dirty_lock); list_move_tail(&ci->i_flushing_item, &cap->session->s_cap_flushing); + oldest_flush_tid = __get_oldest_flush_tid(mdsc); spin_unlock(&mdsc->cap_dirty_lock); + ci->i_ceph_flags &= ~CEPH_I_KICK_FLUSH; + __kick_flushing_caps(mdsc, session, ci, oldest_flush_tid); spin_unlock(&ci->i_ceph_lock); - - delayed = __kick_flushing_caps(mdsc, session, ci); - if (delayed) { - spin_lock(&ci->i_ceph_lock); - __cap_delay_requeue(mdsc, ci); - spin_unlock(&ci->i_ceph_lock); - } } else { spin_unlock(&ci->i_ceph_lock); } @@ -2580,16 +2578,19 @@ void ceph_get_cap_refs(struct ceph_inode_info *ci, int caps) * drop cap_snap that is not associated with any snapshot. * we don't need to send FLUSHSNAP message for it. */ -static int ceph_try_drop_cap_snap(struct ceph_cap_snap *capsnap) +static int ceph_try_drop_cap_snap(struct ceph_inode_info *ci, + struct ceph_cap_snap *capsnap) { if (!capsnap->need_flush && !capsnap->writing && !capsnap->dirty_pages) { - dout("dropping cap_snap %p follows %llu\n", capsnap, capsnap->follows); + BUG_ON(capsnap->cap_flush.tid > 0); ceph_put_snap_context(capsnap->context); + if (!list_is_last(&capsnap->ci_item, &ci->i_cap_snaps)) + ci->i_ceph_flags |= CEPH_I_FLUSH_SNAPS; + list_del(&capsnap->ci_item); - list_del(&capsnap->flushing_item); ceph_put_cap_snap(capsnap); return 1; } @@ -2636,7 +2637,7 @@ void ceph_put_cap_refs(struct ceph_inode_info *ci, int had) struct ceph_cap_snap, ci_item); capsnap->writing = 0; - if (ceph_try_drop_cap_snap(capsnap)) + if (ceph_try_drop_cap_snap(ci, capsnap)) put++; else if (__ceph_finish_cap_snap(ci, capsnap)) flushsnaps = 1; @@ -2661,7 +2662,7 @@ void ceph_put_cap_refs(struct ceph_inode_info *ci, int had) if (last && !flushsnaps) ceph_check_caps(ci, 0, NULL); else if (flushsnaps) - ceph_flush_snaps(ci); + ceph_flush_snaps(ci, NULL); if (wake) wake_up_all(&ci->i_cap_wq); while (put-- > 0) @@ -2679,15 +2680,19 @@ void ceph_put_wrbuffer_cap_refs(struct ceph_inode_info *ci, int nr, struct ceph_snap_context *snapc) { struct inode *inode = &ci->vfs_inode; - int last = 0; - int complete_capsnap = 0; - int drop_capsnap = 0; - int found = 0; struct ceph_cap_snap *capsnap = NULL; + int put = 0; + bool last = false; + bool found = false; + bool flush_snaps = false; + bool complete_capsnap = false; spin_lock(&ci->i_ceph_lock); ci->i_wrbuffer_ref -= nr; - last = !ci->i_wrbuffer_ref; + if (ci->i_wrbuffer_ref == 0) { + last = true; + put++; + } if (ci->i_head_snapc == snapc) { ci->i_wrbuffer_ref_head -= nr; @@ -2707,15 +2712,22 @@ void ceph_put_wrbuffer_cap_refs(struct ceph_inode_info *ci, int nr, } else { list_for_each_entry(capsnap, &ci->i_cap_snaps, ci_item) { if (capsnap->context == snapc) { - found = 1; + found = true; break; } } BUG_ON(!found); capsnap->dirty_pages -= nr; if (capsnap->dirty_pages == 0) { - complete_capsnap = 1; - drop_capsnap = ceph_try_drop_cap_snap(capsnap); + complete_capsnap = true; + if (!capsnap->writing) { + if (ceph_try_drop_cap_snap(ci, capsnap)) { + put++; + } else { + ci->i_ceph_flags |= CEPH_I_FLUSH_SNAPS; + flush_snaps = true; + } + } } dout("put_wrbuffer_cap_refs on %p cap_snap %p " " snap %lld %d/%d -> %d/%d %s%s\n", @@ -2730,12 +2742,12 @@ void ceph_put_wrbuffer_cap_refs(struct ceph_inode_info *ci, int nr, if (last) { ceph_check_caps(ci, CHECK_CAPS_AUTHONLY, NULL); - iput(inode); - } else if (complete_capsnap) { - ceph_flush_snaps(ci); - wake_up_all(&ci->i_cap_wq); + } else if (flush_snaps) { + ceph_flush_snaps(ci, NULL); } - if (drop_capsnap) + if (complete_capsnap) + wake_up_all(&ci->i_cap_wq); + while (put-- > 0) iput(inode); } @@ -2779,12 +2791,11 @@ static void invalidate_aliases(struct inode *inode) */ static void handle_cap_grant(struct ceph_mds_client *mdsc, struct inode *inode, struct ceph_mds_caps *grant, - u64 inline_version, - void *inline_data, int inline_len, + struct ceph_string **pns, u64 inline_version, + void *inline_data, u32 inline_len, struct ceph_buffer *xattr_buf, struct ceph_mds_session *session, - struct ceph_cap *cap, int issued, - u32 pool_ns_len) + struct ceph_cap *cap, int issued) __releases(ci->i_ceph_lock) __releases(mdsc->snap_rwsem) { @@ -2895,8 +2906,18 @@ static void handle_cap_grant(struct ceph_mds_client *mdsc, if (newcaps & (CEPH_CAP_ANY_FILE_RD | CEPH_CAP_ANY_FILE_WR)) { /* file layout may have changed */ - ci->i_layout = grant->layout; - ci->i_pool_ns_len = pool_ns_len; + s64 old_pool = ci->i_layout.pool_id; + struct ceph_string *old_ns; + + ceph_file_layout_from_legacy(&ci->i_layout, &grant->layout); + old_ns = rcu_dereference_protected(ci->i_layout.pool_ns, + lockdep_is_held(&ci->i_ceph_lock)); + rcu_assign_pointer(ci->i_layout.pool_ns, *pns); + + if (ci->i_layout.pool_id != old_pool || *pns != old_ns) + ci->i_ceph_flags &= ~CEPH_I_POOL_PERM; + + *pns = old_ns; /* size/truncate_seq? */ queue_trunc = ceph_fill_file_size(inode, issued, @@ -2979,13 +3000,13 @@ static void handle_cap_grant(struct ceph_mds_client *mdsc, fill_inline = true; } - spin_unlock(&ci->i_ceph_lock); - if (le32_to_cpu(grant->op) == CEPH_CAP_OP_IMPORT) { - kick_flushing_inode_caps(mdsc, session, inode); - up_read(&mdsc->snap_rwsem); if (newcaps & ~issued) wake = true; + kick_flushing_inode_caps(mdsc, session, inode); + up_read(&mdsc->snap_rwsem); + } else { + spin_unlock(&ci->i_ceph_lock); } if (fill_inline) @@ -3029,23 +3050,24 @@ static void handle_cap_flush_ack(struct inode *inode, u64 flush_tid, { struct ceph_inode_info *ci = ceph_inode(inode); struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc; - struct ceph_cap_flush *cf; - struct rb_node *n; + struct ceph_cap_flush *cf, *tmp_cf; LIST_HEAD(to_remove); unsigned seq = le32_to_cpu(m->seq); int dirty = le32_to_cpu(m->dirty); int cleaned = 0; - int drop = 0; + bool drop = false; + bool wake_ci = 0; + bool wake_mdsc = 0; - n = rb_first(&ci->i_cap_flush_tree); - while (n) { - cf = rb_entry(n, struct ceph_cap_flush, i_node); - n = rb_next(&cf->i_node); + list_for_each_entry_safe(cf, tmp_cf, &ci->i_cap_flush_list, i_list) { if (cf->tid == flush_tid) cleaned = cf->caps; + if (cf->caps == 0) /* capsnap */ + continue; if (cf->tid <= flush_tid) { - rb_erase(&cf->i_node, &ci->i_cap_flush_tree); - list_add_tail(&cf->list, &to_remove); + if (__finish_cap_flush(NULL, ci, cf)) + wake_ci = true; + list_add_tail(&cf->i_list, &to_remove); } else { cleaned &= ~cf->caps; if (!cleaned) @@ -3066,31 +3088,29 @@ static void handle_cap_flush_ack(struct inode *inode, u64 flush_tid, spin_lock(&mdsc->cap_dirty_lock); - if (!list_empty(&to_remove)) { - list_for_each_entry(cf, &to_remove, list) - rb_erase(&cf->g_node, &mdsc->cap_flush_tree); - - n = rb_first(&mdsc->cap_flush_tree); - cf = n ? rb_entry(n, struct ceph_cap_flush, g_node) : NULL; - if (!cf || cf->tid > flush_tid) - wake_up_all(&mdsc->cap_flushing_wq); + list_for_each_entry(cf, &to_remove, i_list) { + if (__finish_cap_flush(mdsc, NULL, cf)) + wake_mdsc = true; } if (ci->i_flushing_caps == 0) { - list_del_init(&ci->i_flushing_item); - if (!list_empty(&session->s_cap_flushing)) - dout(" mds%d still flushing cap on %p\n", - session->s_mds, - &list_entry(session->s_cap_flushing.next, - struct ceph_inode_info, - i_flushing_item)->vfs_inode); + if (list_empty(&ci->i_cap_flush_list)) { + list_del_init(&ci->i_flushing_item); + if (!list_empty(&session->s_cap_flushing)) { + dout(" mds%d still flushing cap on %p\n", + session->s_mds, + &list_first_entry(&session->s_cap_flushing, + struct ceph_inode_info, + i_flushing_item)->vfs_inode); + } + } mdsc->num_cap_flushing--; dout(" inode %p now !flushing\n", inode); if (ci->i_dirty_caps == 0) { dout(" inode %p now clean\n", inode); BUG_ON(!list_empty(&ci->i_dirty_item)); - drop = 1; + drop = true; if (ci->i_wr_ref == 0 && ci->i_wrbuffer_ref_head == 0) { BUG_ON(!ci->i_head_snapc); @@ -3102,17 +3122,21 @@ static void handle_cap_flush_ack(struct inode *inode, u64 flush_tid, } } spin_unlock(&mdsc->cap_dirty_lock); - wake_up_all(&ci->i_cap_wq); out: spin_unlock(&ci->i_ceph_lock); while (!list_empty(&to_remove)) { cf = list_first_entry(&to_remove, - struct ceph_cap_flush, list); - list_del(&cf->list); + struct ceph_cap_flush, i_list); + list_del(&cf->i_list); ceph_free_cap_flush(cf); } + + if (wake_ci) + wake_up_all(&ci->i_cap_wq); + if (wake_mdsc) + wake_up_all(&mdsc->cap_flushing_wq); if (drop) iput(inode); } @@ -3131,7 +3155,9 @@ static void handle_cap_flushsnap_ack(struct inode *inode, u64 flush_tid, struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc; u64 follows = le64_to_cpu(m->snap_follows); struct ceph_cap_snap *capsnap; - int drop = 0; + bool flushed = false; + bool wake_ci = false; + bool wake_mdsc = false; dout("handle_cap_flushsnap_ack inode %p ci %p mds%d follows %lld\n", inode, ci, session->s_mds, follows); @@ -3139,30 +3165,47 @@ static void handle_cap_flushsnap_ack(struct inode *inode, u64 flush_tid, spin_lock(&ci->i_ceph_lock); list_for_each_entry(capsnap, &ci->i_cap_snaps, ci_item) { if (capsnap->follows == follows) { - if (capsnap->flush_tid != flush_tid) { + if (capsnap->cap_flush.tid != flush_tid) { dout(" cap_snap %p follows %lld tid %lld !=" " %lld\n", capsnap, follows, - flush_tid, capsnap->flush_tid); + flush_tid, capsnap->cap_flush.tid); break; } - WARN_ON(capsnap->dirty_pages || capsnap->writing); - dout(" removing %p cap_snap %p follows %lld\n", - inode, capsnap, follows); - ceph_put_snap_context(capsnap->context); - list_del(&capsnap->ci_item); - list_del(&capsnap->flushing_item); - ceph_put_cap_snap(capsnap); - wake_up_all(&mdsc->cap_flushing_wq); - drop = 1; + flushed = true; break; } else { dout(" skipping cap_snap %p follows %lld\n", capsnap, capsnap->follows); } } + if (flushed) { + WARN_ON(capsnap->dirty_pages || capsnap->writing); + dout(" removing %p cap_snap %p follows %lld\n", + inode, capsnap, follows); + list_del(&capsnap->ci_item); + if (__finish_cap_flush(NULL, ci, &capsnap->cap_flush)) + wake_ci = true; + + spin_lock(&mdsc->cap_dirty_lock); + + if (list_empty(&ci->i_cap_flush_list)) + list_del_init(&ci->i_flushing_item); + + if (__finish_cap_flush(mdsc, NULL, &capsnap->cap_flush)) + wake_mdsc = true; + + spin_unlock(&mdsc->cap_dirty_lock); + } spin_unlock(&ci->i_ceph_lock); - if (drop) + if (flushed) { + ceph_put_snap_context(capsnap->context); + ceph_put_cap_snap(capsnap); + if (wake_ci) + wake_up_all(&ci->i_cap_wq); + if (wake_mdsc) + wake_up_all(&mdsc->cap_flushing_wq); iput(inode); + } } /* @@ -3267,7 +3310,8 @@ retry: tcap->implemented |= issued; if (cap == ci->i_auth_cap) ci->i_auth_cap = tcap; - if (ci->i_flushing_caps && ci->i_auth_cap == tcap) { + if (!list_empty(&ci->i_cap_flush_list) && + ci->i_auth_cap == tcap) { spin_lock(&mdsc->cap_dirty_lock); list_move_tail(&ci->i_flushing_item, &tcap->session->s_cap_flushing); @@ -3420,20 +3464,18 @@ void ceph_handle_caps(struct ceph_mds_session *session, struct ceph_cap *cap; struct ceph_mds_caps *h; struct ceph_mds_cap_peer *peer = NULL; - struct ceph_snap_realm *realm; + struct ceph_snap_realm *realm = NULL; + struct ceph_string *pool_ns = NULL; int mds = session->s_mds; int op, issued; u32 seq, mseq; struct ceph_vino vino; - u64 cap_id; - u64 size, max_size; u64 tid; u64 inline_version = 0; void *inline_data = NULL; u32 inline_len = 0; void *snaptrace; size_t snaptrace_len; - u32 pool_ns_len = 0; void *p, *end; dout("handle_caps from mds%d\n", mds); @@ -3447,11 +3489,8 @@ void ceph_handle_caps(struct ceph_mds_session *session, op = le32_to_cpu(h->op); vino.ino = le64_to_cpu(h->ino); vino.snap = CEPH_NOSNAP; - cap_id = le64_to_cpu(h->cap_id); seq = le32_to_cpu(h->seq); mseq = le32_to_cpu(h->migrate_seq); - size = le64_to_cpu(h->size); - max_size = le64_to_cpu(h->max_size); snaptrace = h + 1; snaptrace_len = le32_to_cpu(h->snap_trace_len); @@ -3490,6 +3529,7 @@ void ceph_handle_caps(struct ceph_mds_session *session, u64 flush_tid; u32 caller_uid, caller_gid; u32 osd_epoch_barrier; + u32 pool_ns_len; /* version >= 5 */ ceph_decode_32_safe(&p, end, osd_epoch_barrier, bad); /* version >= 6 */ @@ -3499,6 +3539,11 @@ void ceph_handle_caps(struct ceph_mds_session *session, ceph_decode_32_safe(&p, end, caller_gid, bad); /* version >= 8 */ ceph_decode_32_safe(&p, end, pool_ns_len, bad); + if (pool_ns_len > 0) { + ceph_decode_need(&p, end, pool_ns_len, bad); + pool_ns = ceph_find_or_create_string(p, pool_ns_len); + p += pool_ns_len; + } } /* lookup ino */ @@ -3519,7 +3564,7 @@ void ceph_handle_caps(struct ceph_mds_session *session, cap = ceph_get_cap(mdsc, NULL); cap->cap_ino = vino.ino; cap->queue_release = 1; - cap->cap_id = cap_id; + cap->cap_id = le64_to_cpu(h->cap_id); cap->mseq = mseq; cap->seq = seq; spin_lock(&session->s_cap_lock); @@ -3554,10 +3599,9 @@ void ceph_handle_caps(struct ceph_mds_session *session, } handle_cap_import(mdsc, inode, h, peer, session, &cap, &issued); - handle_cap_grant(mdsc, inode, h, + handle_cap_grant(mdsc, inode, h, &pool_ns, inline_version, inline_data, inline_len, - msg->middle, session, cap, issued, - pool_ns_len); + msg->middle, session, cap, issued); if (realm) ceph_put_snap_realm(mdsc, realm); goto done_unlocked; @@ -3579,10 +3623,9 @@ void ceph_handle_caps(struct ceph_mds_session *session, case CEPH_CAP_OP_GRANT: __ceph_caps_issued(ci, &issued); issued |= __ceph_caps_dirty(ci); - handle_cap_grant(mdsc, inode, h, + handle_cap_grant(mdsc, inode, h, &pool_ns, inline_version, inline_data, inline_len, - msg->middle, session, cap, issued, - pool_ns_len); + msg->middle, session, cap, issued); goto done_unlocked; case CEPH_CAP_OP_FLUSH_ACK: @@ -3613,6 +3656,7 @@ done: mutex_unlock(&session->s_mutex); done_unlocked: iput(inode); + ceph_put_string(pool_ns); return; bad: @@ -3673,6 +3717,16 @@ void ceph_flush_dirty_caps(struct ceph_mds_client *mdsc) dout("flush_dirty_caps done\n"); } +void __ceph_get_fmode(struct ceph_inode_info *ci, int fmode) +{ + int i; + int bits = (fmode << 1) | 1; + for (i = 0; i < CEPH_FILE_MODE_BITS; i++) { + if (bits & (1 << i)) + ci->i_nr_by_mode[i]++; + } +} + /* * Drop open file reference. If we were the last open file, * we may need to release capabilities to the MDS (or schedule @@ -3680,15 +3734,20 @@ void ceph_flush_dirty_caps(struct ceph_mds_client *mdsc) */ void ceph_put_fmode(struct ceph_inode_info *ci, int fmode) { - struct inode *inode = &ci->vfs_inode; - int last = 0; - + int i, last = 0; + int bits = (fmode << 1) | 1; spin_lock(&ci->i_ceph_lock); - dout("put_fmode %p fmode %d %d -> %d\n", inode, fmode, - ci->i_nr_by_mode[fmode], ci->i_nr_by_mode[fmode]-1); - BUG_ON(ci->i_nr_by_mode[fmode] == 0); - if (--ci->i_nr_by_mode[fmode] == 0) - last++; + for (i = 0; i < CEPH_FILE_MODE_BITS; i++) { + if (bits & (1 << i)) { + BUG_ON(ci->i_nr_by_mode[i] == 0); + if (--ci->i_nr_by_mode[i] == 0) + last++; + } + } + dout("put_fmode %p fmode %d {%d,%d,%d,%d}\n", + &ci->vfs_inode, fmode, + ci->i_nr_by_mode[0], ci->i_nr_by_mode[1], + ci->i_nr_by_mode[2], ci->i_nr_by_mode[3]); spin_unlock(&ci->i_ceph_lock); if (last && ci->i_vino.snap == CEPH_NOSNAP) diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c index 6e0fedf6713b..c64a0b794d49 100644 --- a/fs/ceph/dir.c +++ b/fs/ceph/dir.c @@ -59,7 +59,7 @@ int ceph_init_dentry(struct dentry *dentry) di->dentry = dentry; di->lease_session = NULL; - dentry->d_time = jiffies; + di->time = jiffies; /* avoid reordering d_fsdata setup so that the check above is safe */ smp_mb(); dentry->d_fsdata = di; @@ -1124,7 +1124,7 @@ static int ceph_rename(struct inode *old_dir, struct dentry *old_dentry, void ceph_invalidate_dentry_lease(struct dentry *dentry) { spin_lock(&dentry->d_lock); - dentry->d_time = jiffies; + ceph_dentry(dentry)->time = jiffies; ceph_dentry(dentry)->lease_shared_gen = 0; spin_unlock(&dentry->d_lock); } @@ -1133,7 +1133,8 @@ void ceph_invalidate_dentry_lease(struct dentry *dentry) * Check if dentry lease is valid. If not, delete the lease. Try to * renew if the least is more than half up. */ -static int dentry_lease_is_valid(struct dentry *dentry) +static int dentry_lease_is_valid(struct dentry *dentry, unsigned int flags, + struct inode *dir) { struct ceph_dentry_info *di; struct ceph_mds_session *s; @@ -1141,12 +1142,11 @@ static int dentry_lease_is_valid(struct dentry *dentry) u32 gen; unsigned long ttl; struct ceph_mds_session *session = NULL; - struct inode *dir = NULL; u32 seq = 0; spin_lock(&dentry->d_lock); di = ceph_dentry(dentry); - if (di->lease_session) { + if (di && di->lease_session) { s = di->lease_session; spin_lock(&s->s_gen_ttl_lock); gen = s->s_cap_gen; @@ -1154,17 +1154,24 @@ static int dentry_lease_is_valid(struct dentry *dentry) spin_unlock(&s->s_gen_ttl_lock); if (di->lease_gen == gen && - time_before(jiffies, dentry->d_time) && + time_before(jiffies, di->time) && time_before(jiffies, ttl)) { valid = 1; if (di->lease_renew_after && time_after(jiffies, di->lease_renew_after)) { - /* we should renew */ - dir = d_inode(dentry->d_parent); - session = ceph_get_mds_session(s); - seq = di->lease_seq; - di->lease_renew_after = 0; - di->lease_renew_from = jiffies; + /* + * We should renew. If we're in RCU walk mode + * though, we can't do that so just return + * -ECHILD. + */ + if (flags & LOOKUP_RCU) { + valid = -ECHILD; + } else { + session = ceph_get_mds_session(s); + seq = di->lease_seq; + di->lease_renew_after = 0; + di->lease_renew_from = jiffies; + } } } } @@ -1207,15 +1214,19 @@ static int ceph_d_revalidate(struct dentry *dentry, unsigned int flags) struct dentry *parent; struct inode *dir; - if (flags & LOOKUP_RCU) - return -ECHILD; + if (flags & LOOKUP_RCU) { + parent = ACCESS_ONCE(dentry->d_parent); + dir = d_inode_rcu(parent); + if (!dir) + return -ECHILD; + } else { + parent = dget_parent(dentry); + dir = d_inode(parent); + } dout("d_revalidate %p '%pd' inode %p offset %lld\n", dentry, dentry, d_inode(dentry), ceph_dentry(dentry)->offset); - parent = dget_parent(dentry); - dir = d_inode(parent); - /* always trust cached snapped dentries, snapdir dentry */ if (ceph_snap(dir) != CEPH_NOSNAP) { dout("d_revalidate %p '%pd' inode %p is SNAPPED\n", dentry, @@ -1224,12 +1235,16 @@ static int ceph_d_revalidate(struct dentry *dentry, unsigned int flags) } else if (d_really_is_positive(dentry) && ceph_snap(d_inode(dentry)) == CEPH_SNAPDIR) { valid = 1; - } else if (dentry_lease_is_valid(dentry) || - dir_lease_is_valid(dir, dentry)) { - if (d_really_is_positive(dentry)) - valid = ceph_is_any_caps(d_inode(dentry)); - else - valid = 1; + } else { + valid = dentry_lease_is_valid(dentry, flags, dir); + if (valid == -ECHILD) + return valid; + if (valid || dir_lease_is_valid(dir, dentry)) { + if (d_really_is_positive(dentry)) + valid = ceph_is_any_caps(d_inode(dentry)); + else + valid = 1; + } } if (!valid) { @@ -1238,6 +1253,9 @@ static int ceph_d_revalidate(struct dentry *dentry, unsigned int flags) struct ceph_mds_request *req; int op, mask, err; + if (flags & LOOKUP_RCU) + return -ECHILD; + op = ceph_snap(dir) == CEPH_SNAPDIR ? CEPH_MDS_OP_LOOKUPSNAP : CEPH_MDS_OP_LOOKUP; req = ceph_mdsc_create_request(mdsc, op, USE_ANY_MDS); @@ -1273,7 +1291,8 @@ static int ceph_d_revalidate(struct dentry *dentry, unsigned int flags) ceph_dir_clear_complete(dir); } - dput(parent); + if (!(flags & LOOKUP_RCU)) + dput(parent); return valid; } @@ -1286,10 +1305,14 @@ static void ceph_d_release(struct dentry *dentry) dout("d_release %p\n", dentry); ceph_dentry_lru_del(dentry); + + spin_lock(&dentry->d_lock); + dentry->d_fsdata = NULL; + spin_unlock(&dentry->d_lock); + if (di->lease_session) ceph_put_mds_session(di->lease_session); kmem_cache_free(ceph_dentry_cachep, di); - dentry->d_fsdata = NULL; } static int ceph_snapdir_d_revalidate(struct dentry *dentry, diff --git a/fs/ceph/file.c b/fs/ceph/file.c index 0daaf7ceedc5..0f5375d8e030 100644 --- a/fs/ceph/file.c +++ b/fs/ceph/file.c @@ -708,7 +708,7 @@ static void ceph_aio_complete_req(struct ceph_osd_request *req) } } - ceph_put_page_vector(osd_data->pages, num_pages, false); + ceph_put_page_vector(osd_data->pages, num_pages, !aio_req->write); ceph_osdc_put_request(req); if (rc < 0) @@ -821,6 +821,54 @@ static void ceph_sync_write_unsafe(struct ceph_osd_request *req, bool unsafe) } } +/* + * Wait on any unsafe replies for the given inode. First wait on the + * newest request, and make that the upper bound. Then, if there are + * more requests, keep waiting on the oldest as long as it is still older + * than the original request. + */ +void ceph_sync_write_wait(struct inode *inode) +{ + struct ceph_inode_info *ci = ceph_inode(inode); + struct list_head *head = &ci->i_unsafe_writes; + struct ceph_osd_request *req; + u64 last_tid; + + if (!S_ISREG(inode->i_mode)) + return; + + spin_lock(&ci->i_unsafe_lock); + if (list_empty(head)) + goto out; + + /* set upper bound as _last_ entry in chain */ + + req = list_last_entry(head, struct ceph_osd_request, + r_unsafe_item); + last_tid = req->r_tid; + + do { + ceph_osdc_get_request(req); + spin_unlock(&ci->i_unsafe_lock); + + dout("sync_write_wait on tid %llu (until %llu)\n", + req->r_tid, last_tid); + wait_for_completion(&req->r_safe_completion); + ceph_osdc_put_request(req); + + spin_lock(&ci->i_unsafe_lock); + /* + * from here on look at first entry in chain, since we + * only want to wait for anything older than last_tid + */ + if (list_empty(head)) + break; + req = list_first_entry(head, struct ceph_osd_request, + r_unsafe_item); + } while (req->r_tid < last_tid); +out: + spin_unlock(&ci->i_unsafe_lock); +} static ssize_t ceph_direct_read_write(struct kiocb *iocb, struct iov_iter *iter, @@ -964,7 +1012,7 @@ ceph_direct_read_write(struct kiocb *iocb, struct iov_iter *iter, len = ret; } - ceph_put_page_vector(pages, num_pages, false); + ceph_put_page_vector(pages, num_pages, !write); ceph_osdc_put_request(req); if (ret < 0) @@ -985,6 +1033,8 @@ ceph_direct_read_write(struct kiocb *iocb, struct iov_iter *iter, } if (aio_req) { + LIST_HEAD(osd_reqs); + if (aio_req->num_reqs == 0) { kfree(aio_req); return ret; @@ -993,8 +1043,9 @@ ceph_direct_read_write(struct kiocb *iocb, struct iov_iter *iter, ceph_get_cap_refs(ci, write ? CEPH_CAP_FILE_WR : CEPH_CAP_FILE_RD); - while (!list_empty(&aio_req->osd_reqs)) { - req = list_first_entry(&aio_req->osd_reqs, + list_splice(&aio_req->osd_reqs, &osd_reqs); + while (!list_empty(&osd_reqs)) { + req = list_first_entry(&osd_reqs, struct ceph_osd_request, r_unsafe_item); list_del_init(&req->r_unsafe_item); @@ -1448,16 +1499,14 @@ static loff_t ceph_llseek(struct file *file, loff_t offset, int whence) { struct inode *inode = file->f_mapping->host; loff_t i_size; - int ret; + loff_t ret; inode_lock(inode); if (whence == SEEK_END || whence == SEEK_DATA || whence == SEEK_HOLE) { ret = ceph_do_getattr(inode, CEPH_STAT_CAP_SIZE, false); - if (ret < 0) { - offset = ret; + if (ret < 0) goto out; - } } i_size = i_size_read(inode); @@ -1473,7 +1522,7 @@ static loff_t ceph_llseek(struct file *file, loff_t offset, int whence) * write() or lseek() might have altered it */ if (offset == 0) { - offset = file->f_pos; + ret = file->f_pos; goto out; } offset += file->f_pos; @@ -1493,11 +1542,11 @@ static loff_t ceph_llseek(struct file *file, loff_t offset, int whence) break; } - offset = vfs_setpos(file, offset, inode->i_sb->s_maxbytes); + ret = vfs_setpos(file, offset, inode->i_sb->s_maxbytes); out: inode_unlock(inode); - return offset; + return ret; } static inline void ceph_zero_partial_page( @@ -1583,9 +1632,9 @@ static int ceph_zero_objects(struct inode *inode, loff_t offset, loff_t length) { int ret = 0; struct ceph_inode_info *ci = ceph_inode(inode); - s32 stripe_unit = ceph_file_layout_su(ci->i_layout); - s32 stripe_count = ceph_file_layout_stripe_count(ci->i_layout); - s32 object_size = ceph_file_layout_object_size(ci->i_layout); + s32 stripe_unit = ci->i_layout.stripe_unit; + s32 stripe_count = ci->i_layout.stripe_count; + s32 object_size = ci->i_layout.object_size; u64 object_set_size = object_size * stripe_count; u64 nearly, t; diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c index 99bdef66213a..dd3a6dbf71eb 100644 --- a/fs/ceph/inode.c +++ b/fs/ceph/inode.c @@ -446,7 +446,7 @@ struct inode *ceph_alloc_inode(struct super_block *sb) ci->i_symlink = NULL; memset(&ci->i_dir_layout, 0, sizeof(ci->i_dir_layout)); - ci->i_pool_ns_len = 0; + RCU_INIT_POINTER(ci->i_layout.pool_ns, NULL); ci->i_fragtree = RB_ROOT; mutex_init(&ci->i_fragtree_mutex); @@ -468,7 +468,7 @@ struct inode *ceph_alloc_inode(struct super_block *sb) INIT_LIST_HEAD(&ci->i_dirty_item); INIT_LIST_HEAD(&ci->i_flushing_item); ci->i_prealloc_cap_flush = NULL; - ci->i_cap_flush_tree = RB_ROOT; + INIT_LIST_HEAD(&ci->i_cap_flush_list); init_waitqueue_head(&ci->i_cap_wq); ci->i_hold_caps_min = 0; ci->i_hold_caps_max = 0; @@ -477,7 +477,7 @@ struct inode *ceph_alloc_inode(struct super_block *sb) ci->i_head_snapc = NULL; ci->i_snap_caps = 0; - for (i = 0; i < CEPH_FILE_MODE_NUM; i++) + for (i = 0; i < CEPH_FILE_MODE_BITS; i++) ci->i_nr_by_mode[i] = 0; mutex_init(&ci->i_truncate_mutex); @@ -570,6 +570,8 @@ void ceph_destroy_inode(struct inode *inode) if (ci->i_xattrs.prealloc_blob) ceph_buffer_put(ci->i_xattrs.prealloc_blob); + ceph_put_string(rcu_dereference_raw(ci->i_layout.pool_ns)); + call_rcu(&inode->i_rcu, ceph_i_callback); } @@ -583,6 +585,14 @@ int ceph_drop_inode(struct inode *inode) return 1; } +void ceph_evict_inode(struct inode *inode) +{ + /* wait unsafe sync writes */ + ceph_sync_write_wait(inode); + truncate_inode_pages_final(&inode->i_data); + clear_inode(inode); +} + static inline blkcnt_t calc_inode_blocks(u64 size) { return (size + (1<<9) - 1) >> 9; @@ -733,6 +743,7 @@ static int fill_inode(struct inode *inode, struct page *locked_page, int issued = 0, implemented, new_issued; struct timespec mtime, atime, ctime; struct ceph_buffer *xattr_blob = NULL; + struct ceph_string *pool_ns = NULL; struct ceph_cap *new_cap = NULL; int err = 0; bool wake = false; @@ -760,6 +771,10 @@ static int fill_inode(struct inode *inode, struct page *locked_page, iinfo->xattr_len); } + if (iinfo->pool_ns_len > 0) + pool_ns = ceph_find_or_create_string(iinfo->pool_ns_data, + iinfo->pool_ns_len); + spin_lock(&ci->i_ceph_lock); /* @@ -814,10 +829,18 @@ static int fill_inode(struct inode *inode, struct page *locked_page, if (new_version || (new_issued & (CEPH_CAP_ANY_FILE_RD | CEPH_CAP_ANY_FILE_WR))) { - if (ci->i_layout.fl_pg_pool != info->layout.fl_pg_pool) + s64 old_pool = ci->i_layout.pool_id; + struct ceph_string *old_ns; + + ceph_file_layout_from_legacy(&ci->i_layout, &info->layout); + old_ns = rcu_dereference_protected(ci->i_layout.pool_ns, + lockdep_is_held(&ci->i_ceph_lock)); + rcu_assign_pointer(ci->i_layout.pool_ns, pool_ns); + + if (ci->i_layout.pool_id != old_pool || pool_ns != old_ns) ci->i_ceph_flags &= ~CEPH_I_POOL_PERM; - ci->i_layout = info->layout; - ci->i_pool_ns_len = iinfo->pool_ns_len; + + pool_ns = old_ns; queue_trunc = ceph_fill_file_size(inode, issued, le32_to_cpu(info->truncate_seq), @@ -985,6 +1008,7 @@ out: ceph_put_cap(mdsc, new_cap); if (xattr_blob) ceph_buffer_put(xattr_blob); + ceph_put_string(pool_ns); return err; } @@ -1018,7 +1042,7 @@ static void update_dentry_lease(struct dentry *dentry, goto out_unlock; if (di->lease_gen == session->s_cap_gen && - time_before(ttl, dentry->d_time)) + time_before(ttl, di->time)) goto out_unlock; /* we already have a newer lease. */ if (di->lease_session && di->lease_session != session) @@ -1032,7 +1056,7 @@ static void update_dentry_lease(struct dentry *dentry, di->lease_seq = le32_to_cpu(lease->seq); di->lease_renew_after = half_ttl; di->lease_renew_from = 0; - dentry->d_time = ttl; + di->time = ttl; out_unlock: spin_unlock(&dentry->d_lock); return; diff --git a/fs/ceph/ioctl.c b/fs/ceph/ioctl.c index be6b1657b1af..7d752d53353a 100644 --- a/fs/ceph/ioctl.c +++ b/fs/ceph/ioctl.c @@ -21,10 +21,10 @@ static long ceph_ioctl_get_layout(struct file *file, void __user *arg) err = ceph_do_getattr(file_inode(file), CEPH_STAT_CAP_LAYOUT, false); if (!err) { - l.stripe_unit = ceph_file_layout_su(ci->i_layout); - l.stripe_count = ceph_file_layout_stripe_count(ci->i_layout); - l.object_size = ceph_file_layout_object_size(ci->i_layout); - l.data_pool = le32_to_cpu(ci->i_layout.fl_pg_pool); + l.stripe_unit = ci->i_layout.stripe_unit; + l.stripe_count = ci->i_layout.stripe_count; + l.object_size = ci->i_layout.object_size; + l.data_pool = ci->i_layout.pool_id; l.preferred_osd = (s32)-1; if (copy_to_user(arg, &l, sizeof(l))) return -EFAULT; @@ -82,19 +82,19 @@ static long ceph_ioctl_set_layout(struct file *file, void __user *arg) if (l.stripe_count) nl.stripe_count = l.stripe_count; else - nl.stripe_count = ceph_file_layout_stripe_count(ci->i_layout); + nl.stripe_count = ci->i_layout.stripe_count; if (l.stripe_unit) nl.stripe_unit = l.stripe_unit; else - nl.stripe_unit = ceph_file_layout_su(ci->i_layout); + nl.stripe_unit = ci->i_layout.stripe_unit; if (l.object_size) nl.object_size = l.object_size; else - nl.object_size = ceph_file_layout_object_size(ci->i_layout); + nl.object_size = ci->i_layout.object_size; if (l.data_pool) nl.data_pool = l.data_pool; else - nl.data_pool = ceph_file_layout_pg_pool(ci->i_layout); + nl.data_pool = ci->i_layout.pool_id; /* this is obsolete, and always -1 */ nl.preferred_osd = le64_to_cpu(-1); @@ -183,7 +183,7 @@ static long ceph_ioctl_get_dataloc(struct file *file, void __user *arg) struct ceph_osd_client *osdc = &ceph_sb_to_client(inode->i_sb)->client->osdc; struct ceph_object_locator oloc; - struct ceph_object_id oid; + CEPH_DEFINE_OID_ONSTACK(oid); u64 len = 1, olen; u64 tmp; struct ceph_pg pgid; @@ -202,8 +202,8 @@ static long ceph_ioctl_get_dataloc(struct file *file, void __user *arg) return -EIO; } dl.file_offset -= dl.object_offset; - dl.object_size = ceph_file_layout_object_size(ci->i_layout); - dl.block_size = ceph_file_layout_su(ci->i_layout); + dl.object_size = ci->i_layout.object_size; + dl.block_size = ci->i_layout.stripe_unit; /* block_offset = object_offset % block_size */ tmp = dl.object_offset; @@ -212,10 +212,13 @@ static long ceph_ioctl_get_dataloc(struct file *file, void __user *arg) snprintf(dl.object_name, sizeof(dl.object_name), "%llx.%08llx", ceph_ino(inode), dl.object_no); - oloc.pool = ceph_file_layout_pg_pool(ci->i_layout); + oloc.pool = ci->i_layout.pool_id; + oloc.pool_ns = ceph_try_get_string(ci->i_layout.pool_ns); ceph_oid_printf(&oid, "%s", dl.object_name); r = ceph_object_locator_to_pg(osdc->osdmap, &oid, &oloc, &pgid); + + ceph_oloc_destroy(&oloc); if (r < 0) { up_read(&osdc->lock); return r; @@ -247,9 +250,8 @@ static long ceph_ioctl_lazyio(struct file *file) if ((fi->fmode & CEPH_FILE_MODE_LAZY) == 0) { spin_lock(&ci->i_ceph_lock); - ci->i_nr_by_mode[fi->fmode]--; fi->fmode |= CEPH_FILE_MODE_LAZY; - ci->i_nr_by_mode[fi->fmode]++; + ci->i_nr_by_mode[ffs(CEPH_FILE_MODE_LAZY)]++; spin_unlock(&ci->i_ceph_lock); dout("ioctl_layzio: file %p marked lazy\n", file); diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c index 4e8678a612b6..fa59a85226b2 100644 --- a/fs/ceph/mds_client.c +++ b/fs/ceph/mds_client.c @@ -48,7 +48,7 @@ struct ceph_reconnect_state { int nr_caps; struct ceph_pagelist *pagelist; - bool flock; + unsigned msg_version; }; static void __wake_requests(struct ceph_mds_client *mdsc, @@ -100,12 +100,15 @@ static int parse_reply_info_in(void **p, void *end, } else info->inline_version = CEPH_INLINE_NONE; + info->pool_ns_len = 0; + info->pool_ns_data = NULL; if (features & CEPH_FEATURE_FS_FILE_LAYOUT_V2) { ceph_decode_32_safe(p, end, info->pool_ns_len, bad); - ceph_decode_need(p, end, info->pool_ns_len, bad); - *p += info->pool_ns_len; - } else { - info->pool_ns_len = 0; + if (info->pool_ns_len > 0) { + ceph_decode_need(p, end, info->pool_ns_len, bad); + info->pool_ns_data = *p; + *p += info->pool_ns_len; + } } return 0; @@ -469,7 +472,6 @@ static struct ceph_mds_session *register_session(struct ceph_mds_client *mdsc, s->s_cap_iterator = NULL; INIT_LIST_HEAD(&s->s_cap_releases); INIT_LIST_HEAD(&s->s_cap_flushing); - INIT_LIST_HEAD(&s->s_cap_snaps_flushing); dout("register_session mds%d\n", mds); if (mds >= mdsc->max_sessions) { @@ -1145,19 +1147,17 @@ static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap, ACCESS_ONCE(fsc->mount_state) == CEPH_MOUNT_SHUTDOWN) invalidate = true; - while (true) { - struct rb_node *n = rb_first(&ci->i_cap_flush_tree); - if (!n) - break; - cf = rb_entry(n, struct ceph_cap_flush, i_node); - rb_erase(&cf->i_node, &ci->i_cap_flush_tree); - list_add(&cf->list, &to_remove); + while (!list_empty(&ci->i_cap_flush_list)) { + cf = list_first_entry(&ci->i_cap_flush_list, + struct ceph_cap_flush, i_list); + list_del(&cf->i_list); + list_add(&cf->i_list, &to_remove); } spin_lock(&mdsc->cap_dirty_lock); - list_for_each_entry(cf, &to_remove, list) - rb_erase(&cf->g_node, &mdsc->cap_flush_tree); + list_for_each_entry(cf, &to_remove, i_list) + list_del(&cf->g_list); if (!list_empty(&ci->i_dirty_item)) { pr_warn_ratelimited( @@ -1181,7 +1181,7 @@ static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap, spin_unlock(&mdsc->cap_dirty_lock); if (!ci->i_dirty_caps && ci->i_prealloc_cap_flush) { - list_add(&ci->i_prealloc_cap_flush->list, &to_remove); + list_add(&ci->i_prealloc_cap_flush->i_list, &to_remove); ci->i_prealloc_cap_flush = NULL; } } @@ -1189,8 +1189,8 @@ static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap, while (!list_empty(&to_remove)) { struct ceph_cap_flush *cf; cf = list_first_entry(&to_remove, - struct ceph_cap_flush, list); - list_del(&cf->list); + struct ceph_cap_flush, i_list); + list_del(&cf->i_list); ceph_free_cap_flush(cf); } @@ -1212,6 +1212,8 @@ static void remove_session_caps(struct ceph_mds_session *session) dout("remove_session_caps on %p\n", session); iterate_session_caps(session, remove_session_caps_cb, fsc); + wake_up_all(&fsc->mdsc->cap_flushing_wq); + spin_lock(&session->s_cap_lock); if (session->s_nr_caps > 0) { struct inode *inode; @@ -1478,35 +1480,21 @@ static int trim_caps(struct ceph_mds_client *mdsc, return 0; } -static int check_capsnap_flush(struct ceph_inode_info *ci, - u64 want_snap_seq) -{ - int ret = 1; - spin_lock(&ci->i_ceph_lock); - if (want_snap_seq > 0 && !list_empty(&ci->i_cap_snaps)) { - struct ceph_cap_snap *capsnap = - list_first_entry(&ci->i_cap_snaps, - struct ceph_cap_snap, ci_item); - ret = capsnap->follows >= want_snap_seq; - } - spin_unlock(&ci->i_ceph_lock); - return ret; -} - static int check_caps_flush(struct ceph_mds_client *mdsc, u64 want_flush_tid) { - struct rb_node *n; - struct ceph_cap_flush *cf; int ret = 1; spin_lock(&mdsc->cap_dirty_lock); - n = rb_first(&mdsc->cap_flush_tree); - cf = n ? rb_entry(n, struct ceph_cap_flush, g_node) : NULL; - if (cf && cf->tid <= want_flush_tid) { - dout("check_caps_flush still flushing tid %llu <= %llu\n", - cf->tid, want_flush_tid); - ret = 0; + if (!list_empty(&mdsc->cap_flush_list)) { + struct ceph_cap_flush *cf = + list_first_entry(&mdsc->cap_flush_list, + struct ceph_cap_flush, g_list); + if (cf->tid <= want_flush_tid) { + dout("check_caps_flush still flushing tid " + "%llu <= %llu\n", cf->tid, want_flush_tid); + ret = 0; + } } spin_unlock(&mdsc->cap_dirty_lock); return ret; @@ -1518,54 +1506,9 @@ static int check_caps_flush(struct ceph_mds_client *mdsc, * returns true if we've flushed through want_flush_tid */ static void wait_caps_flush(struct ceph_mds_client *mdsc, - u64 want_flush_tid, u64 want_snap_seq) + u64 want_flush_tid) { - int mds; - - dout("check_caps_flush want %llu snap want %llu\n", - want_flush_tid, want_snap_seq); - mutex_lock(&mdsc->mutex); - for (mds = 0; mds < mdsc->max_sessions; ) { - struct ceph_mds_session *session = mdsc->sessions[mds]; - struct inode *inode = NULL; - - if (!session) { - mds++; - continue; - } - get_session(session); - mutex_unlock(&mdsc->mutex); - - mutex_lock(&session->s_mutex); - if (!list_empty(&session->s_cap_snaps_flushing)) { - struct ceph_cap_snap *capsnap = - list_first_entry(&session->s_cap_snaps_flushing, - struct ceph_cap_snap, - flushing_item); - struct ceph_inode_info *ci = capsnap->ci; - if (!check_capsnap_flush(ci, want_snap_seq)) { - dout("check_cap_flush still flushing snap %p " - "follows %lld <= %lld to mds%d\n", - &ci->vfs_inode, capsnap->follows, - want_snap_seq, mds); - inode = igrab(&ci->vfs_inode); - } - } - mutex_unlock(&session->s_mutex); - ceph_put_mds_session(session); - - if (inode) { - wait_event(mdsc->cap_flushing_wq, - check_capsnap_flush(ceph_inode(inode), - want_snap_seq)); - iput(inode); - } else { - mds++; - } - - mutex_lock(&mdsc->mutex); - } - mutex_unlock(&mdsc->mutex); + dout("check_caps_flush want %llu\n", want_flush_tid); wait_event(mdsc->cap_flushing_wq, check_caps_flush(mdsc, want_flush_tid)); @@ -2163,6 +2106,11 @@ static int __do_request(struct ceph_mds_client *mdsc, mds = __choose_mds(mdsc, req); if (mds < 0 || ceph_mdsmap_get_state(mdsc->mdsmap, mds) < CEPH_MDS_STATE_ACTIVE) { + if (mdsc->mdsmap_err) { + err = mdsc->mdsmap_err; + dout("do_request mdsmap err %d\n", err); + goto finish; + } dout("do_request no mds or not active, waiting for map\n"); list_add(&req->r_wait, &mdsc->waiting_for_map); goto out; @@ -2292,14 +2240,6 @@ int ceph_mdsc_do_request(struct ceph_mds_client *mdsc, ceph_get_cap_refs(ceph_inode(req->r_old_dentry_dir), CEPH_CAP_PIN); - /* deny access to directories with pool_ns layouts */ - if (req->r_inode && S_ISDIR(req->r_inode->i_mode) && - ceph_inode(req->r_inode)->i_pool_ns_len) - return -EIO; - if (req->r_locked_dir && - ceph_inode(req->r_locked_dir)->i_pool_ns_len) - return -EIO; - /* issue */ mutex_lock(&mdsc->mutex); __register_request(mdsc, req, dir); @@ -2791,13 +2731,13 @@ static int encode_caps_cb(struct inode *inode, struct ceph_cap *cap, struct ceph_mds_cap_reconnect v2; struct ceph_mds_cap_reconnect_v1 v1; } rec; - size_t reclen; struct ceph_inode_info *ci; struct ceph_reconnect_state *recon_state = arg; struct ceph_pagelist *pagelist = recon_state->pagelist; char *path; int pathlen, err; u64 pathbase; + u64 snap_follows; struct dentry *dentry; ci = cap->ci; @@ -2820,9 +2760,6 @@ static int encode_caps_cb(struct inode *inode, struct ceph_cap *cap, path = NULL; pathlen = 0; } - err = ceph_pagelist_encode_string(pagelist, path, pathlen); - if (err) - goto out_free; spin_lock(&ci->i_ceph_lock); cap->seq = 0; /* reset cap seq */ @@ -2830,14 +2767,13 @@ static int encode_caps_cb(struct inode *inode, struct ceph_cap *cap, cap->mseq = 0; /* and migrate_seq */ cap->cap_gen = cap->session->s_cap_gen; - if (recon_state->flock) { + if (recon_state->msg_version >= 2) { rec.v2.cap_id = cpu_to_le64(cap->cap_id); rec.v2.wanted = cpu_to_le32(__ceph_caps_wanted(ci)); rec.v2.issued = cpu_to_le32(cap->issued); rec.v2.snaprealm = cpu_to_le64(ci->i_snap_realm->ino); rec.v2.pathbase = cpu_to_le64(pathbase); rec.v2.flock_len = 0; - reclen = sizeof(rec.v2); } else { rec.v1.cap_id = cpu_to_le64(cap->cap_id); rec.v1.wanted = cpu_to_le32(__ceph_caps_wanted(ci)); @@ -2847,13 +2783,23 @@ static int encode_caps_cb(struct inode *inode, struct ceph_cap *cap, ceph_encode_timespec(&rec.v1.atime, &inode->i_atime); rec.v1.snaprealm = cpu_to_le64(ci->i_snap_realm->ino); rec.v1.pathbase = cpu_to_le64(pathbase); - reclen = sizeof(rec.v1); + } + + if (list_empty(&ci->i_cap_snaps)) { + snap_follows = 0; + } else { + struct ceph_cap_snap *capsnap = + list_first_entry(&ci->i_cap_snaps, + struct ceph_cap_snap, ci_item); + snap_follows = capsnap->follows; } spin_unlock(&ci->i_ceph_lock); - if (recon_state->flock) { + if (recon_state->msg_version >= 2) { int num_fcntl_locks, num_flock_locks; struct ceph_filelock *flocks; + size_t struct_len, total_len = 0; + u8 struct_v = 0; encode_again: ceph_count_locks(inode, &num_fcntl_locks, &num_flock_locks); @@ -2872,20 +2818,51 @@ encode_again: goto encode_again; goto out_free; } + + if (recon_state->msg_version >= 3) { + /* version, compat_version and struct_len */ + total_len = 2 * sizeof(u8) + sizeof(u32); + struct_v = 2; + } /* * number of encoded locks is stable, so copy to pagelist */ - rec.v2.flock_len = cpu_to_le32(2*sizeof(u32) + - (num_fcntl_locks+num_flock_locks) * - sizeof(struct ceph_filelock)); - err = ceph_pagelist_append(pagelist, &rec, reclen); - if (!err) - err = ceph_locks_to_pagelist(flocks, pagelist, - num_fcntl_locks, - num_flock_locks); + struct_len = 2 * sizeof(u32) + + (num_fcntl_locks + num_flock_locks) * + sizeof(struct ceph_filelock); + rec.v2.flock_len = cpu_to_le32(struct_len); + + struct_len += sizeof(rec.v2); + struct_len += sizeof(u32) + pathlen; + + if (struct_v >= 2) + struct_len += sizeof(u64); /* snap_follows */ + + total_len += struct_len; + err = ceph_pagelist_reserve(pagelist, total_len); + + if (!err) { + if (recon_state->msg_version >= 3) { + ceph_pagelist_encode_8(pagelist, struct_v); + ceph_pagelist_encode_8(pagelist, 1); + ceph_pagelist_encode_32(pagelist, struct_len); + } + ceph_pagelist_encode_string(pagelist, path, pathlen); + ceph_pagelist_append(pagelist, &rec, sizeof(rec.v2)); + ceph_locks_to_pagelist(flocks, pagelist, + num_fcntl_locks, + num_flock_locks); + if (struct_v >= 2) + ceph_pagelist_encode_64(pagelist, snap_follows); + } kfree(flocks); } else { - err = ceph_pagelist_append(pagelist, &rec, reclen); + size_t size = sizeof(u32) + pathlen + sizeof(rec.v1); + err = ceph_pagelist_reserve(pagelist, size); + if (!err) { + ceph_pagelist_encode_string(pagelist, path, pathlen); + ceph_pagelist_append(pagelist, &rec, sizeof(rec.v1)); + } } recon_state->nr_caps++; @@ -2976,7 +2953,12 @@ static void send_mds_reconnect(struct ceph_mds_client *mdsc, recon_state.nr_caps = 0; recon_state.pagelist = pagelist; - recon_state.flock = session->s_con.peer_features & CEPH_FEATURE_FLOCK; + if (session->s_con.peer_features & CEPH_FEATURE_MDSENC) + recon_state.msg_version = 3; + else if (session->s_con.peer_features & CEPH_FEATURE_FLOCK) + recon_state.msg_version = 2; + else + recon_state.msg_version = 1; err = iterate_session_caps(session, encode_caps_cb, &recon_state); if (err < 0) goto fail; @@ -3005,8 +2987,7 @@ static void send_mds_reconnect(struct ceph_mds_client *mdsc, goto fail; } - if (recon_state.flock) - reply->hdr.version = cpu_to_le16(2); + reply->hdr.version = cpu_to_le16(recon_state.msg_version); /* raced with cap release? */ if (s_nr_caps != recon_state.nr_caps) { @@ -3231,7 +3212,7 @@ static void handle_lease(struct ceph_mds_client *mdsc, msecs_to_jiffies(le32_to_cpu(h->duration_ms)); di->lease_seq = seq; - dentry->d_time = di->lease_renew_from + duration; + di->time = di->lease_renew_from + duration; di->lease_renew_after = di->lease_renew_from + (duration >> 1); di->lease_renew_from = 0; @@ -3297,47 +3278,6 @@ void ceph_mdsc_lease_send_msg(struct ceph_mds_session *session, } /* - * Preemptively release a lease we expect to invalidate anyway. - * Pass @inode always, @dentry is optional. - */ -void ceph_mdsc_lease_release(struct ceph_mds_client *mdsc, struct inode *inode, - struct dentry *dentry) -{ - struct ceph_dentry_info *di; - struct ceph_mds_session *session; - u32 seq; - - BUG_ON(inode == NULL); - BUG_ON(dentry == NULL); - - /* is dentry lease valid? */ - spin_lock(&dentry->d_lock); - di = ceph_dentry(dentry); - if (!di || !di->lease_session || - di->lease_session->s_mds < 0 || - di->lease_gen != di->lease_session->s_cap_gen || - !time_before(jiffies, dentry->d_time)) { - dout("lease_release inode %p dentry %p -- " - "no lease\n", - inode, dentry); - spin_unlock(&dentry->d_lock); - return; - } - - /* we do have a lease on this dentry; note mds and seq */ - session = ceph_get_mds_session(di->lease_session); - seq = di->lease_seq; - __ceph_mdsc_drop_dentry_lease(dentry); - spin_unlock(&dentry->d_lock); - - dout("lease_release inode %p dentry %p to mds%d\n", - inode, dentry, session->s_mds); - ceph_mdsc_lease_send_msg(session, inode, dentry, - CEPH_MDS_LEASE_RELEASE, seq); - ceph_put_mds_session(session); -} - -/* * drop all leases (and dentry refs) in preparation for umount */ static void drop_leases(struct ceph_mds_client *mdsc) @@ -3470,7 +3410,7 @@ int ceph_mdsc_init(struct ceph_fs_client *fsc) INIT_LIST_HEAD(&mdsc->snap_flush_list); spin_lock_init(&mdsc->snap_flush_lock); mdsc->last_cap_flush_tid = 1; - mdsc->cap_flush_tree = RB_ROOT; + INIT_LIST_HEAD(&mdsc->cap_flush_list); INIT_LIST_HEAD(&mdsc->cap_dirty); INIT_LIST_HEAD(&mdsc->cap_dirty_migrating); mdsc->num_cap_flushing = 0; @@ -3585,7 +3525,7 @@ restart: void ceph_mdsc_sync(struct ceph_mds_client *mdsc) { - u64 want_tid, want_flush, want_snap; + u64 want_tid, want_flush; if (ACCESS_ONCE(mdsc->fsc->mount_state) == CEPH_MOUNT_SHUTDOWN) return; @@ -3598,17 +3538,19 @@ void ceph_mdsc_sync(struct ceph_mds_client *mdsc) ceph_flush_dirty_caps(mdsc); spin_lock(&mdsc->cap_dirty_lock); want_flush = mdsc->last_cap_flush_tid; + if (!list_empty(&mdsc->cap_flush_list)) { + struct ceph_cap_flush *cf = + list_last_entry(&mdsc->cap_flush_list, + struct ceph_cap_flush, g_list); + cf->wake = true; + } spin_unlock(&mdsc->cap_dirty_lock); - down_read(&mdsc->snap_rwsem); - want_snap = mdsc->last_snap_seq; - up_read(&mdsc->snap_rwsem); - - dout("sync want tid %lld flush_seq %lld snap_seq %lld\n", - want_tid, want_flush, want_snap); + dout("sync want tid %lld flush_seq %lld\n", + want_tid, want_flush); wait_unsafe_requests(mdsc, want_tid); - wait_caps_flush(mdsc, want_flush, want_snap); + wait_caps_flush(mdsc, want_flush); } /* @@ -3729,11 +3671,86 @@ void ceph_mdsc_destroy(struct ceph_fs_client *fsc) dout("mdsc_destroy %p done\n", mdsc); } +void ceph_mdsc_handle_fsmap(struct ceph_mds_client *mdsc, struct ceph_msg *msg) +{ + struct ceph_fs_client *fsc = mdsc->fsc; + const char *mds_namespace = fsc->mount_options->mds_namespace; + void *p = msg->front.iov_base; + void *end = p + msg->front.iov_len; + u32 epoch; + u32 map_len; + u32 num_fs; + u32 mount_fscid = (u32)-1; + u8 struct_v, struct_cv; + int err = -EINVAL; + + ceph_decode_need(&p, end, sizeof(u32), bad); + epoch = ceph_decode_32(&p); + + dout("handle_fsmap epoch %u\n", epoch); + + ceph_decode_need(&p, end, 2 + sizeof(u32), bad); + struct_v = ceph_decode_8(&p); + struct_cv = ceph_decode_8(&p); + map_len = ceph_decode_32(&p); + + ceph_decode_need(&p, end, sizeof(u32) * 3, bad); + p += sizeof(u32) * 2; /* skip epoch and legacy_client_fscid */ + + num_fs = ceph_decode_32(&p); + while (num_fs-- > 0) { + void *info_p, *info_end; + u32 info_len; + u8 info_v, info_cv; + u32 fscid, namelen; + + ceph_decode_need(&p, end, 2 + sizeof(u32), bad); + info_v = ceph_decode_8(&p); + info_cv = ceph_decode_8(&p); + info_len = ceph_decode_32(&p); + ceph_decode_need(&p, end, info_len, bad); + info_p = p; + info_end = p + info_len; + p = info_end; + + ceph_decode_need(&info_p, info_end, sizeof(u32) * 2, bad); + fscid = ceph_decode_32(&info_p); + namelen = ceph_decode_32(&info_p); + ceph_decode_need(&info_p, info_end, namelen, bad); + + if (mds_namespace && + strlen(mds_namespace) == namelen && + !strncmp(mds_namespace, (char *)info_p, namelen)) { + mount_fscid = fscid; + break; + } + } + + ceph_monc_got_map(&fsc->client->monc, CEPH_SUB_FSMAP, epoch); + if (mount_fscid != (u32)-1) { + fsc->client->monc.fs_cluster_id = mount_fscid; + ceph_monc_want_map(&fsc->client->monc, CEPH_SUB_MDSMAP, + 0, true); + ceph_monc_renew_subs(&fsc->client->monc); + } else { + err = -ENOENT; + goto err_out; + } + return; +bad: + pr_err("error decoding fsmap\n"); +err_out: + mutex_lock(&mdsc->mutex); + mdsc->mdsmap_err = -ENOENT; + __wake_requests(mdsc, &mdsc->waiting_for_map); + mutex_unlock(&mdsc->mutex); + return; +} /* * handle mds map update. */ -void ceph_mdsc_handle_map(struct ceph_mds_client *mdsc, struct ceph_msg *msg) +void ceph_mdsc_handle_mdsmap(struct ceph_mds_client *mdsc, struct ceph_msg *msg) { u32 epoch; u32 maplen; @@ -3840,7 +3857,10 @@ static void dispatch(struct ceph_connection *con, struct ceph_msg *msg) switch (type) { case CEPH_MSG_MDS_MAP: - ceph_mdsc_handle_map(mdsc, msg); + ceph_mdsc_handle_mdsmap(mdsc, msg); + break; + case CEPH_MSG_FS_MAP_USER: + ceph_mdsc_handle_fsmap(mdsc, msg); break; case CEPH_MSG_CLIENT_SESSION: handle_session(s, msg); diff --git a/fs/ceph/mds_client.h b/fs/ceph/mds_client.h index e7d38aac7109..6b3679737d4a 100644 --- a/fs/ceph/mds_client.h +++ b/fs/ceph/mds_client.h @@ -45,6 +45,7 @@ struct ceph_mds_reply_info_in { u32 inline_len; char *inline_data; u32 pool_ns_len; + char *pool_ns_data; }; struct ceph_mds_reply_dir_entry { @@ -151,7 +152,6 @@ struct ceph_mds_session { /* protected by mutex */ struct list_head s_cap_flushing; /* inodes w/ flushing caps */ - struct list_head s_cap_snaps_flushing; unsigned long s_renew_requested; /* last time we sent a renew req */ u64 s_renew_seq; @@ -275,8 +275,10 @@ struct ceph_mds_request { struct ceph_pool_perm { struct rb_node node; - u32 pool; int perm; + s64 pool; + size_t pool_ns_len; + char pool_ns[]; }; /* @@ -290,6 +292,7 @@ struct ceph_mds_client { struct completion safe_umount_waiters; wait_queue_head_t session_close_wq; struct list_head waiting_for_map; + int mdsmap_err; struct ceph_mds_session **sessions; /* NULL for mds if no session */ atomic_t num_sessions; @@ -321,7 +324,7 @@ struct ceph_mds_client { spinlock_t snap_flush_lock; u64 last_cap_flush_tid; - struct rb_root cap_flush_tree; + struct list_head cap_flush_list; struct list_head cap_dirty; /* inodes with dirty caps */ struct list_head cap_dirty_migrating; /* ...that are migration... */ int num_cap_flushing; /* # caps we are flushing */ @@ -382,10 +385,6 @@ extern void ceph_mdsc_destroy(struct ceph_fs_client *fsc); extern void ceph_mdsc_sync(struct ceph_mds_client *mdsc); -extern void ceph_mdsc_lease_release(struct ceph_mds_client *mdsc, - struct inode *inode, - struct dentry *dn); - extern void ceph_invalidate_dir_request(struct ceph_mds_request *req); extern int ceph_alloc_readdir_reply_buffer(struct ceph_mds_request *req, struct inode *dir); @@ -420,8 +419,10 @@ extern void ceph_mdsc_lease_send_msg(struct ceph_mds_session *session, struct dentry *dentry, char action, u32 seq); -extern void ceph_mdsc_handle_map(struct ceph_mds_client *mdsc, - struct ceph_msg *msg); +extern void ceph_mdsc_handle_mdsmap(struct ceph_mds_client *mdsc, + struct ceph_msg *msg); +extern void ceph_mdsc_handle_fsmap(struct ceph_mds_client *mdsc, + struct ceph_msg *msg); extern struct ceph_mds_session * ceph_mdsc_open_export_target_session(struct ceph_mds_client *mdsc, int target); diff --git a/fs/ceph/snap.c b/fs/ceph/snap.c index 9caaa7ffc93f..9ff5219d849e 100644 --- a/fs/ceph/snap.c +++ b/fs/ceph/snap.c @@ -520,9 +520,7 @@ void ceph_queue_cap_snap(struct ceph_inode_info *ci) ihold(inode); atomic_set(&capsnap->nref, 1); - capsnap->ci = ci; INIT_LIST_HEAD(&capsnap->ci_item); - INIT_LIST_HEAD(&capsnap->flushing_item); capsnap->follows = old_snapc->seq; capsnap->issued = __ceph_caps_issued(ci, NULL); @@ -551,7 +549,6 @@ void ceph_queue_cap_snap(struct ceph_inode_info *ci) ci->i_wrbuffer_ref_head = 0; capsnap->context = old_snapc; list_add_tail(&capsnap->ci_item, &ci->i_cap_snaps); - old_snapc = NULL; if (used & CEPH_CAP_FILE_WR) { dout("queue_cap_snap %p cap_snap %p snapc %p" @@ -563,6 +560,7 @@ void ceph_queue_cap_snap(struct ceph_inode_info *ci) __ceph_finish_cap_snap(ci, capsnap); } capsnap = NULL; + old_snapc = NULL; update_snapc: if (ci->i_head_snapc) { @@ -603,6 +601,8 @@ int __ceph_finish_cap_snap(struct ceph_inode_info *ci, capsnap->dirty_pages); return 0; } + + ci->i_ceph_flags |= CEPH_I_FLUSH_SNAPS; dout("finish_cap_snap %p cap_snap %p snapc %p %llu %s s=%llu\n", inode, capsnap, capsnap->context, capsnap->context->seq, ceph_cap_string(capsnap->dirty), @@ -799,9 +799,7 @@ static void flush_snaps(struct ceph_mds_client *mdsc) inode = &ci->vfs_inode; ihold(inode); spin_unlock(&mdsc->snap_flush_lock); - spin_lock(&ci->i_ceph_lock); - __ceph_flush_snaps(ci, &session, 0); - spin_unlock(&ci->i_ceph_lock); + ceph_flush_snaps(ci, &session); iput(inode); spin_lock(&mdsc->snap_flush_lock); } diff --git a/fs/ceph/super.c b/fs/ceph/super.c index 91e02481ce06..e247f6f0feb7 100644 --- a/fs/ceph/super.c +++ b/fs/ceph/super.c @@ -108,7 +108,6 @@ static int ceph_sync_fs(struct super_block *sb, int wait) * mount options */ enum { - Opt_mds_namespace, Opt_wsize, Opt_rsize, Opt_rasize, @@ -121,6 +120,7 @@ enum { Opt_last_int, /* int args above */ Opt_snapdirname, + Opt_mds_namespace, Opt_last_string, /* string args above */ Opt_dirstat, @@ -144,7 +144,6 @@ enum { }; static match_table_t fsopt_tokens = { - {Opt_mds_namespace, "mds_namespace=%d"}, {Opt_wsize, "wsize=%d"}, {Opt_rsize, "rsize=%d"}, {Opt_rasize, "rasize=%d"}, @@ -156,6 +155,7 @@ static match_table_t fsopt_tokens = { {Opt_congestion_kb, "write_congestion_kb=%d"}, /* int args above */ {Opt_snapdirname, "snapdirname=%s"}, + {Opt_mds_namespace, "mds_namespace=%s"}, /* string args above */ {Opt_dirstat, "dirstat"}, {Opt_nodirstat, "nodirstat"}, @@ -212,11 +212,14 @@ static int parse_fsopt_token(char *c, void *private) if (!fsopt->snapdir_name) return -ENOMEM; break; - - /* misc */ case Opt_mds_namespace: - fsopt->mds_namespace = intval; + fsopt->mds_namespace = kstrndup(argstr[0].from, + argstr[0].to-argstr[0].from, + GFP_KERNEL); + if (!fsopt->mds_namespace) + return -ENOMEM; break; + /* misc */ case Opt_wsize: fsopt->wsize = intval; break; @@ -302,6 +305,7 @@ static void destroy_mount_options(struct ceph_mount_options *args) { dout("destroy_mount_options %p\n", args); kfree(args->snapdir_name); + kfree(args->mds_namespace); kfree(args->server_path); kfree(args); } @@ -333,6 +337,9 @@ static int compare_mount_options(struct ceph_mount_options *new_fsopt, ret = strcmp_null(fsopt1->snapdir_name, fsopt2->snapdir_name); if (ret) return ret; + ret = strcmp_null(fsopt1->mds_namespace, fsopt2->mds_namespace); + if (ret) + return ret; ret = strcmp_null(fsopt1->server_path, fsopt2->server_path); if (ret) @@ -376,7 +383,6 @@ static int parse_mount_options(struct ceph_mount_options **pfsopt, fsopt->max_readdir = CEPH_MAX_READDIR_DEFAULT; fsopt->max_readdir_bytes = CEPH_MAX_READDIR_BYTES_DEFAULT; fsopt->congestion_kb = default_congestion_kb(); - fsopt->mds_namespace = CEPH_FS_CLUSTER_ID_NONE; /* * Distinguish the server list from the path in "dev_name". @@ -469,8 +475,8 @@ static int ceph_show_options(struct seq_file *m, struct dentry *root) seq_puts(m, ",noacl"); #endif - if (fsopt->mds_namespace != CEPH_FS_CLUSTER_ID_NONE) - seq_printf(m, ",mds_namespace=%d", fsopt->mds_namespace); + if (fsopt->mds_namespace) + seq_printf(m, ",mds_namespace=%s", fsopt->mds_namespace); if (fsopt->wsize) seq_printf(m, ",wsize=%d", fsopt->wsize); if (fsopt->rsize != CEPH_RSIZE_DEFAULT) @@ -509,9 +515,11 @@ static int extra_mon_dispatch(struct ceph_client *client, struct ceph_msg *msg) switch (type) { case CEPH_MSG_MDS_MAP: - ceph_mdsc_handle_map(fsc->mdsc, msg); + ceph_mdsc_handle_mdsmap(fsc->mdsc, msg); + return 0; + case CEPH_MSG_FS_MAP_USER: + ceph_mdsc_handle_fsmap(fsc->mdsc, msg); return 0; - default: return -1; } @@ -543,8 +551,14 @@ static struct ceph_fs_client *create_fs_client(struct ceph_mount_options *fsopt, goto fail; } fsc->client->extra_mon_dispatch = extra_mon_dispatch; - fsc->client->monc.fs_cluster_id = fsopt->mds_namespace; - ceph_monc_want_map(&fsc->client->monc, CEPH_SUB_MDSMAP, 0, true); + + if (fsopt->mds_namespace == NULL) { + ceph_monc_want_map(&fsc->client->monc, CEPH_SUB_MDSMAP, + 0, true); + } else { + ceph_monc_want_map(&fsc->client->monc, CEPH_SUB_FSMAP, + 0, false); + } fsc->mount_options = fsopt; @@ -672,8 +686,8 @@ static int __init init_caches(void) if (ceph_dentry_cachep == NULL) goto bad_dentry; - ceph_file_cachep = KMEM_CACHE(ceph_file_info, - SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD); + ceph_file_cachep = KMEM_CACHE(ceph_file_info, SLAB_MEM_SPREAD); + if (ceph_file_cachep == NULL) goto bad_file; @@ -731,6 +745,7 @@ static const struct super_operations ceph_super_ops = { .destroy_inode = ceph_destroy_inode, .write_inode = ceph_write_inode, .drop_inode = ceph_drop_inode, + .evict_inode = ceph_evict_inode, .sync_fs = ceph_sync_fs, .put_super = ceph_put_super, .show_options = ceph_show_options, diff --git a/fs/ceph/super.h b/fs/ceph/super.h index 0168b49fb6ad..3e3fa9163059 100644 --- a/fs/ceph/super.h +++ b/fs/ceph/super.h @@ -62,7 +62,6 @@ struct ceph_mount_options { int cap_release_safety; int max_readdir; /* max readdir result (entires) */ int max_readdir_bytes; /* max readdir result (bytes) */ - int mds_namespace; /* * everything above this point can be memcmp'd; everything below @@ -70,6 +69,7 @@ struct ceph_mount_options { */ char *snapdir_name; /* default ".snap" */ + char *mds_namespace; /* default NULL */ char *server_path; /* default "/" */ }; @@ -147,6 +147,14 @@ struct ceph_cap { #define CHECK_CAPS_AUTHONLY 2 /* only check auth cap */ #define CHECK_CAPS_FLUSH 4 /* flush any dirty caps */ +struct ceph_cap_flush { + u64 tid; + int caps; /* 0 means capsnap */ + bool wake; /* wake up flush waiters when finish ? */ + struct list_head g_list; // global + struct list_head i_list; // per inode +}; + /* * Snapped cap state that is pending flush to mds. When a snapshot occurs, * we first complete any in-process sync writes and writeback any dirty @@ -154,10 +162,11 @@ struct ceph_cap { */ struct ceph_cap_snap { atomic_t nref; - struct ceph_inode_info *ci; - struct list_head ci_item, flushing_item; + struct list_head ci_item; + + struct ceph_cap_flush cap_flush; - u64 follows, flush_tid; + u64 follows; int issued, dirty; struct ceph_snap_context *context; @@ -186,16 +195,6 @@ static inline void ceph_put_cap_snap(struct ceph_cap_snap *capsnap) } } -struct ceph_cap_flush { - u64 tid; - int caps; - struct rb_node g_node; // global - union { - struct rb_node i_node; // inode - struct list_head list; - }; -}; - /* * The frag tree describes how a directory is fragmented, potentially across * multiple metadata servers. It is also used to indicate points where @@ -246,7 +245,7 @@ struct ceph_dentry_info { unsigned long lease_renew_after, lease_renew_from; struct list_head lru; struct dentry *dentry; - u64 time; + unsigned long time; u64 offset; }; @@ -287,7 +286,6 @@ struct ceph_inode_info { struct ceph_dir_layout i_dir_layout; struct ceph_file_layout i_layout; - size_t i_pool_ns_len; char *i_symlink; /* for dirs */ @@ -311,7 +309,7 @@ struct ceph_inode_info { * overlapping, pipelined cap flushes to the mds. we can probably * reduce the tid to 8 bits if we're concerned about inode size. */ struct ceph_cap_flush *i_prealloc_cap_flush; - struct rb_root i_cap_flush_tree; + struct list_head i_cap_flush_list; wait_queue_head_t i_cap_wq; /* threads waiting on a capability */ unsigned long i_hold_caps_min; /* jiffies */ unsigned long i_hold_caps_max; /* jiffies */ @@ -322,7 +320,7 @@ struct ceph_inode_info { dirty|flushing caps */ unsigned i_snap_caps; /* cap bits for snapped files */ - int i_nr_by_mode[CEPH_FILE_MODE_NUM]; /* open file counts */ + int i_nr_by_mode[CEPH_FILE_MODE_BITS]; /* open file counts */ struct mutex i_truncate_mutex; u32 i_truncate_seq; /* last truncate to smaller size */ @@ -471,6 +469,8 @@ static inline struct inode *ceph_find_inode(struct super_block *sb, #define CEPH_I_POOL_WR (1 << 6) /* can write to pool */ #define CEPH_I_SEC_INITED (1 << 7) /* security initialized */ #define CEPH_I_CAP_DROPPED (1 << 8) /* caps were forcibly dropped */ +#define CEPH_I_KICK_FLUSH (1 << 9) /* kick flushing caps */ +#define CEPH_I_FLUSH_SNAPS (1 << 10) /* need flush snapss */ static inline void __ceph_dir_set_complete(struct ceph_inode_info *ci, long long release_count, @@ -750,6 +750,7 @@ extern const struct inode_operations ceph_file_iops; extern struct inode *ceph_alloc_inode(struct super_block *sb); extern void ceph_destroy_inode(struct inode *inode); extern int ceph_drop_inode(struct inode *inode); +extern void ceph_evict_inode(struct inode *inode); extern struct inode *ceph_get_inode(struct super_block *sb, struct ceph_vino vino); @@ -890,9 +891,8 @@ extern void ceph_get_cap_refs(struct ceph_inode_info *ci, int caps); extern void ceph_put_cap_refs(struct ceph_inode_info *ci, int had); extern void ceph_put_wrbuffer_cap_refs(struct ceph_inode_info *ci, int nr, struct ceph_snap_context *snapc); -extern void __ceph_flush_snaps(struct ceph_inode_info *ci, - struct ceph_mds_session **psession, - int again); +extern void ceph_flush_snaps(struct ceph_inode_info *ci, + struct ceph_mds_session **psession); extern void ceph_check_caps(struct ceph_inode_info *ci, int flags, struct ceph_mds_session *session); extern void ceph_check_delayed_caps(struct ceph_mds_client *mdsc); @@ -907,10 +907,7 @@ extern int ceph_get_caps(struct ceph_inode_info *ci, int need, int want, loff_t endoff, int *got, struct page **pinned_page); /* for counting open files by mode */ -static inline void __ceph_get_fmode(struct ceph_inode_info *ci, int mode) -{ - ci->i_nr_by_mode[mode]++; -} +extern void __ceph_get_fmode(struct ceph_inode_info *ci, int mode); extern void ceph_put_fmode(struct ceph_inode_info *ci, int mode); /* addr.c */ @@ -931,6 +928,7 @@ extern int ceph_atomic_open(struct inode *dir, struct dentry *dentry, extern int ceph_release(struct inode *inode, struct file *filp); extern void ceph_fill_inline_data(struct inode *inode, struct page *locked_page, char *data, size_t len); +extern void ceph_sync_write_wait(struct inode *inode); /* dir.c */ extern const struct file_operations ceph_dir_fops; extern const struct file_operations ceph_snapdir_fops; diff --git a/fs/ceph/xattr.c b/fs/ceph/xattr.c index 4870b29df224..adc231892b0d 100644 --- a/fs/ceph/xattr.c +++ b/fs/ceph/xattr.c @@ -57,81 +57,88 @@ struct ceph_vxattr { static bool ceph_vxattrcb_layout_exists(struct ceph_inode_info *ci) { - size_t s; - char *p = (char *)&ci->i_layout; - - for (s = 0; s < sizeof(ci->i_layout); s++, p++) - if (*p) - return true; - return false; + struct ceph_file_layout *fl = &ci->i_layout; + return (fl->stripe_unit > 0 || fl->stripe_count > 0 || + fl->object_size > 0 || fl->pool_id >= 0 || + rcu_dereference_raw(fl->pool_ns) != NULL); } static size_t ceph_vxattrcb_layout(struct ceph_inode_info *ci, char *val, size_t size) { - int ret; struct ceph_fs_client *fsc = ceph_sb_to_client(ci->vfs_inode.i_sb); struct ceph_osd_client *osdc = &fsc->client->osdc; - s64 pool = ceph_file_layout_pg_pool(ci->i_layout); + struct ceph_string *pool_ns; + s64 pool = ci->i_layout.pool_id; const char *pool_name; + const char *ns_field = " pool_namespace="; char buf[128]; + size_t len, total_len = 0; + int ret; + + pool_ns = ceph_try_get_string(ci->i_layout.pool_ns); dout("ceph_vxattrcb_layout %p\n", &ci->vfs_inode); down_read(&osdc->lock); pool_name = ceph_pg_pool_name_by_id(osdc->osdmap, pool); if (pool_name) { - size_t len = strlen(pool_name); - ret = snprintf(buf, sizeof(buf), - "stripe_unit=%lld stripe_count=%lld object_size=%lld pool=", - (unsigned long long)ceph_file_layout_su(ci->i_layout), - (unsigned long long)ceph_file_layout_stripe_count(ci->i_layout), - (unsigned long long)ceph_file_layout_object_size(ci->i_layout)); - if (!size) { - ret += len; - } else if (ret + len > size) { - ret = -ERANGE; - } else { - memcpy(val, buf, ret); + len = snprintf(buf, sizeof(buf), + "stripe_unit=%u stripe_count=%u object_size=%u pool=", + ci->i_layout.stripe_unit, ci->i_layout.stripe_count, + ci->i_layout.object_size); + total_len = len + strlen(pool_name); + } else { + len = snprintf(buf, sizeof(buf), + "stripe_unit=%u stripe_count=%u object_size=%u pool=%lld", + ci->i_layout.stripe_unit, ci->i_layout.stripe_count, + ci->i_layout.object_size, (unsigned long long)pool); + total_len = len; + } + + if (pool_ns) + total_len += strlen(ns_field) + pool_ns->len; + + if (!size) { + ret = total_len; + } else if (total_len > size) { + ret = -ERANGE; + } else { + memcpy(val, buf, len); + ret = len; + if (pool_name) { + len = strlen(pool_name); memcpy(val + ret, pool_name, len); ret += len; } - } else { - ret = snprintf(buf, sizeof(buf), - "stripe_unit=%lld stripe_count=%lld object_size=%lld pool=%lld", - (unsigned long long)ceph_file_layout_su(ci->i_layout), - (unsigned long long)ceph_file_layout_stripe_count(ci->i_layout), - (unsigned long long)ceph_file_layout_object_size(ci->i_layout), - (unsigned long long)pool); - if (size) { - if (ret <= size) - memcpy(val, buf, ret); - else - ret = -ERANGE; + if (pool_ns) { + len = strlen(ns_field); + memcpy(val + ret, ns_field, len); + ret += len; + memcpy(val + ret, pool_ns->str, pool_ns->len); + ret += pool_ns->len; } } up_read(&osdc->lock); + ceph_put_string(pool_ns); return ret; } static size_t ceph_vxattrcb_layout_stripe_unit(struct ceph_inode_info *ci, char *val, size_t size) { - return snprintf(val, size, "%lld", - (unsigned long long)ceph_file_layout_su(ci->i_layout)); + return snprintf(val, size, "%u", ci->i_layout.stripe_unit); } static size_t ceph_vxattrcb_layout_stripe_count(struct ceph_inode_info *ci, char *val, size_t size) { - return snprintf(val, size, "%lld", - (unsigned long long)ceph_file_layout_stripe_count(ci->i_layout)); + return snprintf(val, size, "%u", ci->i_layout.stripe_count); } static size_t ceph_vxattrcb_layout_object_size(struct ceph_inode_info *ci, char *val, size_t size) { - return snprintf(val, size, "%lld", - (unsigned long long)ceph_file_layout_object_size(ci->i_layout)); + return snprintf(val, size, "%u", ci->i_layout.object_size); } static size_t ceph_vxattrcb_layout_pool(struct ceph_inode_info *ci, @@ -140,7 +147,7 @@ static size_t ceph_vxattrcb_layout_pool(struct ceph_inode_info *ci, int ret; struct ceph_fs_client *fsc = ceph_sb_to_client(ci->vfs_inode.i_sb); struct ceph_osd_client *osdc = &fsc->client->osdc; - s64 pool = ceph_file_layout_pg_pool(ci->i_layout); + s64 pool = ci->i_layout.pool_id; const char *pool_name; down_read(&osdc->lock); @@ -153,6 +160,18 @@ static size_t ceph_vxattrcb_layout_pool(struct ceph_inode_info *ci, return ret; } +static size_t ceph_vxattrcb_layout_pool_namespace(struct ceph_inode_info *ci, + char *val, size_t size) +{ + int ret = 0; + struct ceph_string *ns = ceph_try_get_string(ci->i_layout.pool_ns); + if (ns) { + ret = snprintf(val, size, "%.*s", (int)ns->len, ns->str); + ceph_put_string(ns); + } + return ret; +} + /* directories */ static size_t ceph_vxattrcb_dir_entries(struct ceph_inode_info *ci, char *val, @@ -241,6 +260,7 @@ static struct ceph_vxattr ceph_dir_vxattrs[] = { XATTR_LAYOUT_FIELD(dir, layout, stripe_count), XATTR_LAYOUT_FIELD(dir, layout, object_size), XATTR_LAYOUT_FIELD(dir, layout, pool), + XATTR_LAYOUT_FIELD(dir, layout, pool_namespace), XATTR_NAME_CEPH(dir, entries), XATTR_NAME_CEPH(dir, files), XATTR_NAME_CEPH(dir, subdirs), @@ -268,6 +288,7 @@ static struct ceph_vxattr ceph_file_vxattrs[] = { XATTR_LAYOUT_FIELD(file, layout, stripe_count), XATTR_LAYOUT_FIELD(file, layout, object_size), XATTR_LAYOUT_FIELD(file, layout, pool), + XATTR_LAYOUT_FIELD(file, layout, pool_namespace), { .name = NULL, 0 } /* Required table terminator */ }; static size_t ceph_file_vxattrs_name_size; /* total size of all names */ diff --git a/fs/exec.c b/fs/exec.c index ca239fc86d8d..a1789cd684bf 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -866,7 +866,8 @@ int kernel_read_file(struct file *file, void **buf, loff_t *size, goto out; } - *buf = vmalloc(i_size); + if (id != READING_FIRMWARE_PREALLOC_BUFFER) + *buf = vmalloc(i_size); if (!*buf) { ret = -ENOMEM; goto out; @@ -897,8 +898,10 @@ int kernel_read_file(struct file *file, void **buf, loff_t *size, out_free: if (ret < 0) { - vfree(*buf); - *buf = NULL; + if (id != READING_FIRMWARE_PREALLOC_BUFFER) { + vfree(*buf); + *buf = NULL; + } } out: diff --git a/fs/inode.c b/fs/inode.c index 9cef4e16aeda..ad445542c285 100644 --- a/fs/inode.c +++ b/fs/inode.c @@ -345,7 +345,7 @@ EXPORT_SYMBOL(inc_nlink); void address_space_init_once(struct address_space *mapping) { memset(mapping, 0, sizeof(*mapping)); - INIT_RADIX_TREE(&mapping->page_tree, GFP_ATOMIC); + INIT_RADIX_TREE(&mapping->page_tree, GFP_ATOMIC | __GFP_ACCOUNT); spin_lock_init(&mapping->tree_lock); init_rwsem(&mapping->i_mmap_rwsem); INIT_LIST_HEAD(&mapping->private_list); diff --git a/fs/nfs/Makefile b/fs/nfs/Makefile index 8664417955a2..6abdda209642 100644 --- a/fs/nfs/Makefile +++ b/fs/nfs/Makefile @@ -6,7 +6,7 @@ obj-$(CONFIG_NFS_FS) += nfs.o CFLAGS_nfstrace.o += -I$(src) nfs-y := client.o dir.o file.o getroot.o inode.o super.o \ - direct.o pagelist.o read.o symlink.o unlink.o \ + io.o direct.o pagelist.o read.o symlink.o unlink.o \ write.o namespace.o mount_clnt.o nfstrace.o nfs-$(CONFIG_ROOT_NFS) += nfsroot.o nfs-$(CONFIG_SYSCTL) += sysctl.o diff --git a/fs/nfs/blocklayout/dev.c b/fs/nfs/blocklayout/dev.c index e5b89675263e..a69ef4e9c24c 100644 --- a/fs/nfs/blocklayout/dev.c +++ b/fs/nfs/blocklayout/dev.c @@ -65,8 +65,8 @@ nfs4_block_decode_volume(struct xdr_stream *xdr, struct pnfs_block_volume *b) if (!p) return -EIO; b->simple.nr_sigs = be32_to_cpup(p++); - if (!b->simple.nr_sigs) { - dprintk("no signature\n"); + if (!b->simple.nr_sigs || b->simple.nr_sigs > PNFS_BLOCK_MAX_UUIDS) { + dprintk("Bad signature count: %d\n", b->simple.nr_sigs); return -EIO; } @@ -89,7 +89,8 @@ nfs4_block_decode_volume(struct xdr_stream *xdr, struct pnfs_block_volume *b) memcpy(&b->simple.sigs[i].sig, p, b->simple.sigs[i].sig_len); - b->simple.len += 8 + 4 + b->simple.sigs[i].sig_len; + b->simple.len += 8 + 4 + \ + (XDR_QUADLEN(b->simple.sigs[i].sig_len) << 2); } break; case PNFS_BLOCK_VOLUME_SLICE: @@ -104,7 +105,12 @@ nfs4_block_decode_volume(struct xdr_stream *xdr, struct pnfs_block_volume *b) p = xdr_inline_decode(xdr, 4); if (!p) return -EIO; + b->concat.volumes_count = be32_to_cpup(p++); + if (b->concat.volumes_count > PNFS_BLOCK_MAX_DEVICES) { + dprintk("Too many volumes: %d\n", b->concat.volumes_count); + return -EIO; + } p = xdr_inline_decode(xdr, b->concat.volumes_count * 4); if (!p) @@ -116,8 +122,13 @@ nfs4_block_decode_volume(struct xdr_stream *xdr, struct pnfs_block_volume *b) p = xdr_inline_decode(xdr, 8 + 4); if (!p) return -EIO; + p = xdr_decode_hyper(p, &b->stripe.chunk_size); b->stripe.volumes_count = be32_to_cpup(p++); + if (b->stripe.volumes_count > PNFS_BLOCK_MAX_DEVICES) { + dprintk("Too many volumes: %d\n", b->stripe.volumes_count); + return -EIO; + } p = xdr_inline_decode(xdr, b->stripe.volumes_count * 4); if (!p) @@ -224,18 +235,20 @@ bl_parse_simple(struct nfs_server *server, struct pnfs_block_dev *d, struct pnfs_block_volume *volumes, int idx, gfp_t gfp_mask) { struct pnfs_block_volume *v = &volumes[idx]; + struct block_device *bdev; dev_t dev; dev = bl_resolve_deviceid(server, v, gfp_mask); if (!dev) return -EIO; - d->bdev = blkdev_get_by_dev(dev, FMODE_READ | FMODE_WRITE, NULL); - if (IS_ERR(d->bdev)) { + bdev = blkdev_get_by_dev(dev, FMODE_READ | FMODE_WRITE, NULL); + if (IS_ERR(bdev)) { printk(KERN_WARNING "pNFS: failed to open device %d:%d (%ld)\n", - MAJOR(dev), MINOR(dev), PTR_ERR(d->bdev)); - return PTR_ERR(d->bdev); + MAJOR(dev), MINOR(dev), PTR_ERR(bdev)); + return PTR_ERR(bdev); } + d->bdev = bdev; d->len = i_size_read(d->bdev->bd_inode); @@ -287,44 +300,71 @@ bl_validate_designator(struct pnfs_block_volume *v) } } +/* + * Try to open the udev path for the WWN. At least on Debian the udev + * by-id path will always point to the dm-multipath device if one exists. + */ +static struct block_device * +bl_open_udev_path(struct pnfs_block_volume *v) +{ + struct block_device *bdev; + const char *devname; + + devname = kasprintf(GFP_KERNEL, "/dev/disk/by-id/wwn-0x%*phN", + v->scsi.designator_len, v->scsi.designator); + if (!devname) + return ERR_PTR(-ENOMEM); + + bdev = blkdev_get_by_path(devname, FMODE_READ | FMODE_WRITE, NULL); + if (IS_ERR(bdev)) { + pr_warn("pNFS: failed to open device %s (%ld)\n", + devname, PTR_ERR(bdev)); + } + + kfree(devname); + return bdev; +} + +/* + * Try to open the RH/Fedora specific dm-mpath udev path for this WWN, as the + * wwn- links will only point to the first discovered SCSI device there. + */ +static struct block_device * +bl_open_dm_mpath_udev_path(struct pnfs_block_volume *v) +{ + struct block_device *bdev; + const char *devname; + + devname = kasprintf(GFP_KERNEL, + "/dev/disk/by-id/dm-uuid-mpath-%d%*phN", + v->scsi.designator_type, + v->scsi.designator_len, v->scsi.designator); + if (!devname) + return ERR_PTR(-ENOMEM); + + bdev = blkdev_get_by_path(devname, FMODE_READ | FMODE_WRITE, NULL); + kfree(devname); + return bdev; +} + static int bl_parse_scsi(struct nfs_server *server, struct pnfs_block_dev *d, struct pnfs_block_volume *volumes, int idx, gfp_t gfp_mask) { struct pnfs_block_volume *v = &volumes[idx]; + struct block_device *bdev; const struct pr_ops *ops; - const char *devname; int error; if (!bl_validate_designator(v)) return -EINVAL; - switch (v->scsi.designator_len) { - case 8: - devname = kasprintf(GFP_KERNEL, "/dev/disk/by-id/wwn-0x%8phN", - v->scsi.designator); - break; - case 12: - devname = kasprintf(GFP_KERNEL, "/dev/disk/by-id/wwn-0x%12phN", - v->scsi.designator); - break; - case 16: - devname = kasprintf(GFP_KERNEL, "/dev/disk/by-id/wwn-0x%16phN", - v->scsi.designator); - break; - default: - return -EINVAL; - } - - d->bdev = blkdev_get_by_path(devname, FMODE_READ, NULL); - if (IS_ERR(d->bdev)) { - pr_warn("pNFS: failed to open device %s (%ld)\n", - devname, PTR_ERR(d->bdev)); - kfree(devname); - return PTR_ERR(d->bdev); - } - - kfree(devname); + bdev = bl_open_dm_mpath_udev_path(v); + if (IS_ERR(bdev)) + bdev = bl_open_udev_path(v); + if (IS_ERR(bdev)) + return PTR_ERR(bdev); + d->bdev = bdev; d->len = i_size_read(d->bdev->bd_inode); d->map = bl_map_simple; @@ -352,7 +392,7 @@ bl_parse_scsi(struct nfs_server *server, struct pnfs_block_dev *d, return 0; out_blkdev_put: - blkdev_put(d->bdev, FMODE_READ); + blkdev_put(d->bdev, FMODE_READ | FMODE_WRITE); return error; } diff --git a/fs/nfs/blocklayout/extent_tree.c b/fs/nfs/blocklayout/extent_tree.c index 720b3ff55fa9..992bcb19c11e 100644 --- a/fs/nfs/blocklayout/extent_tree.c +++ b/fs/nfs/blocklayout/extent_tree.c @@ -121,6 +121,16 @@ ext_try_to_merge_right(struct rb_root *root, struct pnfs_block_extent *be) return be; } +static void __ext_put_deviceids(struct list_head *head) +{ + struct pnfs_block_extent *be, *tmp; + + list_for_each_entry_safe(be, tmp, head, be_list) { + nfs4_put_deviceid_node(be->be_device); + kfree(be); + } +} + static void __ext_tree_insert(struct rb_root *root, struct pnfs_block_extent *new, bool merge_ok) @@ -163,7 +173,8 @@ free_new: } static int -__ext_tree_remove(struct rb_root *root, sector_t start, sector_t end) +__ext_tree_remove(struct rb_root *root, + sector_t start, sector_t end, struct list_head *tmp) { struct pnfs_block_extent *be; sector_t len1 = 0, len2 = 0; @@ -223,8 +234,7 @@ __ext_tree_remove(struct rb_root *root, sector_t start, sector_t end) struct pnfs_block_extent *next = ext_tree_next(be); rb_erase(&be->be_node, root); - nfs4_put_deviceid_node(be->be_device); - kfree(be); + list_add_tail(&be->be_list, tmp); be = next; } @@ -350,16 +360,18 @@ int ext_tree_remove(struct pnfs_block_layout *bl, bool rw, sector_t start, sector_t end) { int err, err2; + LIST_HEAD(tmp); spin_lock(&bl->bl_ext_lock); - err = __ext_tree_remove(&bl->bl_ext_ro, start, end); + err = __ext_tree_remove(&bl->bl_ext_ro, start, end, &tmp); if (rw) { - err2 = __ext_tree_remove(&bl->bl_ext_rw, start, end); + err2 = __ext_tree_remove(&bl->bl_ext_rw, start, end, &tmp); if (!err) err = err2; } spin_unlock(&bl->bl_ext_lock); + __ext_put_deviceids(&tmp); return err; } @@ -396,12 +408,13 @@ ext_tree_mark_written(struct pnfs_block_layout *bl, sector_t start, sector_t end = start + len; struct pnfs_block_extent *be; int err = 0; + LIST_HEAD(tmp); spin_lock(&bl->bl_ext_lock); /* * First remove all COW extents or holes from written to range. */ - err = __ext_tree_remove(&bl->bl_ext_ro, start, end); + err = __ext_tree_remove(&bl->bl_ext_ro, start, end, &tmp); if (err) goto out; @@ -459,6 +472,8 @@ ext_tree_mark_written(struct pnfs_block_layout *bl, sector_t start, } out: spin_unlock(&bl->bl_ext_lock); + + __ext_put_deviceids(&tmp); return err; } diff --git a/fs/nfs/callback_proc.c b/fs/nfs/callback_proc.c index aaa2e8d3df6f..c92a75e066a6 100644 --- a/fs/nfs/callback_proc.c +++ b/fs/nfs/callback_proc.c @@ -119,27 +119,30 @@ out: * hashed by filehandle. */ static struct pnfs_layout_hdr * get_layout_by_fh_locked(struct nfs_client *clp, - struct nfs_fh *fh, nfs4_stateid *stateid) + struct nfs_fh *fh) { struct nfs_server *server; + struct nfs_inode *nfsi; struct inode *ino; struct pnfs_layout_hdr *lo; +restart: list_for_each_entry_rcu(server, &clp->cl_superblocks, client_link) { list_for_each_entry(lo, &server->layouts, plh_layouts) { - if (!nfs4_stateid_match_other(&lo->plh_stateid, stateid)) + nfsi = NFS_I(lo->plh_inode); + if (nfs_compare_fh(fh, &nfsi->fh)) continue; - if (nfs_compare_fh(fh, &NFS_I(lo->plh_inode)->fh)) + if (nfsi->layout != lo) continue; ino = igrab(lo->plh_inode); if (!ino) break; spin_lock(&ino->i_lock); /* Is this layout in the process of being freed? */ - if (NFS_I(ino)->layout != lo) { + if (nfsi->layout != lo) { spin_unlock(&ino->i_lock); iput(ino); - break; + goto restart; } pnfs_get_layout_hdr(lo); spin_unlock(&ino->i_lock); @@ -151,13 +154,13 @@ static struct pnfs_layout_hdr * get_layout_by_fh_locked(struct nfs_client *clp, } static struct pnfs_layout_hdr * get_layout_by_fh(struct nfs_client *clp, - struct nfs_fh *fh, nfs4_stateid *stateid) + struct nfs_fh *fh) { struct pnfs_layout_hdr *lo; spin_lock(&clp->cl_lock); rcu_read_lock(); - lo = get_layout_by_fh_locked(clp, fh, stateid); + lo = get_layout_by_fh_locked(clp, fh); rcu_read_unlock(); spin_unlock(&clp->cl_lock); @@ -167,17 +170,39 @@ static struct pnfs_layout_hdr * get_layout_by_fh(struct nfs_client *clp, /* * Enforce RFC5661 section 12.5.5.2.1. (Layout Recall and Return Sequencing) */ -static bool pnfs_check_stateid_sequence(struct pnfs_layout_hdr *lo, +static u32 pnfs_check_callback_stateid(struct pnfs_layout_hdr *lo, const nfs4_stateid *new) { u32 oldseq, newseq; - oldseq = be32_to_cpu(lo->plh_stateid.seqid); + /* Is the stateid still not initialised? */ + if (!pnfs_layout_is_valid(lo)) + return NFS4ERR_DELAY; + + /* Mismatched stateid? */ + if (!nfs4_stateid_match_other(&lo->plh_stateid, new)) + return NFS4ERR_BAD_STATEID; + newseq = be32_to_cpu(new->seqid); + /* Are we already in a layout recall situation? */ + if (test_bit(NFS_LAYOUT_RETURN_REQUESTED, &lo->plh_flags) && + lo->plh_return_seq != 0) { + if (newseq < lo->plh_return_seq) + return NFS4ERR_OLD_STATEID; + if (newseq > lo->plh_return_seq) + return NFS4ERR_DELAY; + goto out; + } + /* Check that the stateid matches what we think it should be. */ + oldseq = be32_to_cpu(lo->plh_stateid.seqid); if (newseq > oldseq + 1) - return false; - return true; + return NFS4ERR_DELAY; + /* Crazy server! */ + if (newseq <= oldseq) + return NFS4ERR_OLD_STATEID; +out: + return NFS_OK; } static u32 initiate_file_draining(struct nfs_client *clp, @@ -188,7 +213,7 @@ static u32 initiate_file_draining(struct nfs_client *clp, u32 rv = NFS4ERR_NOMATCHING_LAYOUT; LIST_HEAD(free_me_list); - lo = get_layout_by_fh(clp, &args->cbl_fh, &args->cbl_stateid); + lo = get_layout_by_fh(clp, &args->cbl_fh); if (!lo) { trace_nfs4_cb_layoutrecall_file(clp, &args->cbl_fh, NULL, &args->cbl_stateid, -rv); @@ -196,18 +221,15 @@ static u32 initiate_file_draining(struct nfs_client *clp, } ino = lo->plh_inode; + pnfs_layoutcommit_inode(ino, false); + spin_lock(&ino->i_lock); - if (!pnfs_check_stateid_sequence(lo, &args->cbl_stateid)) { - rv = NFS4ERR_DELAY; + rv = pnfs_check_callback_stateid(lo, &args->cbl_stateid); + if (rv != NFS_OK) goto unlock; - } pnfs_set_layout_stateid(lo, &args->cbl_stateid, true); - spin_unlock(&ino->i_lock); - - pnfs_layoutcommit_inode(ino, false); - spin_lock(&ino->i_lock); /* * Enforce RFC5661 Section 12.5.5.2.1.5 (Bulk Recall and Return) */ @@ -223,11 +245,13 @@ static u32 initiate_file_draining(struct nfs_client *clp, goto unlock; } + /* Embrace your forgetfulness! */ + rv = NFS4ERR_NOMATCHING_LAYOUT; + if (NFS_SERVER(ino)->pnfs_curr_ld->return_range) { NFS_SERVER(ino)->pnfs_curr_ld->return_range(lo, &args->cbl_range); } - pnfs_mark_layout_returned_if_empty(lo); unlock: spin_unlock(&ino->i_lock); pnfs_free_lseg_list(&free_me_list); diff --git a/fs/nfs/callback_xdr.c b/fs/nfs/callback_xdr.c index d81f96aacd51..656f68f7fe53 100644 --- a/fs/nfs/callback_xdr.c +++ b/fs/nfs/callback_xdr.c @@ -925,7 +925,7 @@ static __be32 nfs4_callback_compound(struct svc_rqst *rqstp, void *argp, void *r if (hdr_arg.minorversion == 0) { cps.clp = nfs4_find_client_ident(SVC_NET(rqstp), hdr_arg.cb_ident); if (!cps.clp || !check_gss_callback_principal(cps.clp, rqstp)) - return rpc_drop_reply; + goto out_invalidcred; } cps.minorversion = hdr_arg.minorversion; @@ -953,6 +953,10 @@ static __be32 nfs4_callback_compound(struct svc_rqst *rqstp, void *argp, void *r nfs_put_client(cps.clp); dprintk("%s: done, status = %u\n", __func__, ntohl(status)); return rpc_success; + +out_invalidcred: + pr_warn_ratelimited("NFS: NFSv4 callback contains invalid cred\n"); + return rpc_autherr_badcred; } /* diff --git a/fs/nfs/client.c b/fs/nfs/client.c index 487c5607d52f..003ebce4bbc4 100644 --- a/fs/nfs/client.c +++ b/fs/nfs/client.c @@ -367,8 +367,6 @@ nfs_found_client(const struct nfs_client_initdata *cl_init, */ struct nfs_client * nfs_get_client(const struct nfs_client_initdata *cl_init, - const struct rpc_timeout *timeparms, - const char *ip_addr, rpc_authflavor_t authflavour) { struct nfs_client *clp, *new = NULL; @@ -399,7 +397,7 @@ nfs_get_client(const struct nfs_client_initdata *cl_init, &nn->nfs_client_list); spin_unlock(&nn->nfs_client_lock); new->cl_flags = cl_init->init_flags; - return rpc_ops->init_client(new, timeparms, ip_addr); + return rpc_ops->init_client(new, cl_init); } spin_unlock(&nn->nfs_client_lock); @@ -470,7 +468,7 @@ EXPORT_SYMBOL_GPL(nfs_init_timeout_values); * Create an RPC client handle */ int nfs_create_rpc_client(struct nfs_client *clp, - const struct rpc_timeout *timeparms, + const struct nfs_client_initdata *cl_init, rpc_authflavor_t flavor) { struct rpc_clnt *clnt = NULL; @@ -479,8 +477,9 @@ int nfs_create_rpc_client(struct nfs_client *clp, .protocol = clp->cl_proto, .address = (struct sockaddr *)&clp->cl_addr, .addrsize = clp->cl_addrlen, - .timeout = timeparms, + .timeout = cl_init->timeparms, .servername = clp->cl_hostname, + .nodename = cl_init->nodename, .program = &nfs_program, .version = clp->rpc_ops->version, .authflavor = flavor, @@ -591,14 +590,12 @@ EXPORT_SYMBOL_GPL(nfs_init_server_rpcclient); * nfs_init_client - Initialise an NFS2 or NFS3 client * * @clp: nfs_client to initialise - * @timeparms: timeout parameters for underlying RPC transport - * @ip_addr: IP presentation address (not used) + * @cl_init: Initialisation parameters * * Returns pointer to an NFS client, or an ERR_PTR value. */ struct nfs_client *nfs_init_client(struct nfs_client *clp, - const struct rpc_timeout *timeparms, - const char *ip_addr) + const struct nfs_client_initdata *cl_init) { int error; @@ -612,7 +609,7 @@ struct nfs_client *nfs_init_client(struct nfs_client *clp, * Create a client RPC handle for doing FSSTAT with UNIX auth only * - RFC 2623, sec 2.3.2 */ - error = nfs_create_rpc_client(clp, timeparms, RPC_AUTH_UNIX); + error = nfs_create_rpc_client(clp, cl_init, RPC_AUTH_UNIX); if (error < 0) goto error; nfs_mark_client_ready(clp, NFS_CS_READY); @@ -633,6 +630,7 @@ static int nfs_init_server(struct nfs_server *server, const struct nfs_parsed_mount_data *data, struct nfs_subversion *nfs_mod) { + struct rpc_timeout timeparms; struct nfs_client_initdata cl_init = { .hostname = data->nfs_server.hostname, .addr = (const struct sockaddr *)&data->nfs_server.address, @@ -640,8 +638,8 @@ static int nfs_init_server(struct nfs_server *server, .nfs_mod = nfs_mod, .proto = data->nfs_server.protocol, .net = data->net, + .timeparms = &timeparms, }; - struct rpc_timeout timeparms; struct nfs_client *clp; int error; @@ -653,7 +651,7 @@ static int nfs_init_server(struct nfs_server *server, set_bit(NFS_CS_NORESVPORT, &cl_init.init_flags); /* Allocate or find a client reference we can use */ - clp = nfs_get_client(&cl_init, &timeparms, NULL, RPC_AUTH_UNIX); + clp = nfs_get_client(&cl_init, RPC_AUTH_UNIX); if (IS_ERR(clp)) { dprintk("<-- nfs_init_server() = error %ld\n", PTR_ERR(clp)); return PTR_ERR(clp); diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c index baaa38859899..177fefb26c18 100644 --- a/fs/nfs/dir.c +++ b/fs/nfs/dir.c @@ -2252,21 +2252,37 @@ static struct nfs_access_entry *nfs_access_search_rbtree(struct inode *inode, st return NULL; } -static int nfs_access_get_cached(struct inode *inode, struct rpc_cred *cred, struct nfs_access_entry *res) +static int nfs_access_get_cached(struct inode *inode, struct rpc_cred *cred, struct nfs_access_entry *res, bool may_block) { struct nfs_inode *nfsi = NFS_I(inode); struct nfs_access_entry *cache; - int err = -ENOENT; + bool retry = true; + int err; spin_lock(&inode->i_lock); - if (nfsi->cache_validity & NFS_INO_INVALID_ACCESS) - goto out_zap; - cache = nfs_access_search_rbtree(inode, cred); - if (cache == NULL) - goto out; - if (!nfs_have_delegated_attributes(inode) && - !time_in_range_open(jiffies, cache->jiffies, cache->jiffies + nfsi->attrtimeo)) - goto out_stale; + for(;;) { + if (nfsi->cache_validity & NFS_INO_INVALID_ACCESS) + goto out_zap; + cache = nfs_access_search_rbtree(inode, cred); + err = -ENOENT; + if (cache == NULL) + goto out; + /* Found an entry, is our attribute cache valid? */ + if (!nfs_attribute_cache_expired(inode) && + !(nfsi->cache_validity & NFS_INO_INVALID_ATTR)) + break; + err = -ECHILD; + if (!may_block) + goto out; + if (!retry) + goto out_zap; + spin_unlock(&inode->i_lock); + err = __nfs_revalidate_inode(NFS_SERVER(inode), inode); + if (err) + return err; + spin_lock(&inode->i_lock); + retry = false; + } res->jiffies = cache->jiffies; res->cred = cache->cred; res->mask = cache->mask; @@ -2275,12 +2291,6 @@ static int nfs_access_get_cached(struct inode *inode, struct rpc_cred *cred, str out: spin_unlock(&inode->i_lock); return err; -out_stale: - rb_erase(&cache->rb_node, &nfsi->access_cache); - list_del(&cache->lru); - spin_unlock(&inode->i_lock); - nfs_access_free_entry(cache); - return -ENOENT; out_zap: spin_unlock(&inode->i_lock); nfs_access_zap_cache(inode); @@ -2307,13 +2317,12 @@ static int nfs_access_get_cached_rcu(struct inode *inode, struct rpc_cred *cred, cache = NULL; if (cache == NULL) goto out; - if (!nfs_have_delegated_attributes(inode) && - !time_in_range_open(jiffies, cache->jiffies, cache->jiffies + nfsi->attrtimeo)) + err = nfs_revalidate_inode_rcu(NFS_SERVER(inode), inode); + if (err) goto out; res->jiffies = cache->jiffies; res->cred = cache->cred; res->mask = cache->mask; - err = 0; out: rcu_read_unlock(); return err; @@ -2402,18 +2411,19 @@ EXPORT_SYMBOL_GPL(nfs_access_set_mask); static int nfs_do_access(struct inode *inode, struct rpc_cred *cred, int mask) { struct nfs_access_entry cache; + bool may_block = (mask & MAY_NOT_BLOCK) == 0; int status; trace_nfs_access_enter(inode); status = nfs_access_get_cached_rcu(inode, cred, &cache); if (status != 0) - status = nfs_access_get_cached(inode, cred, &cache); + status = nfs_access_get_cached(inode, cred, &cache, may_block); if (status == 0) goto out_cached; status = -ECHILD; - if (mask & MAY_NOT_BLOCK) + if (!may_block) goto out; /* Be clever: ask server to check for all possible rights */ diff --git a/fs/nfs/direct.c b/fs/nfs/direct.c index e6210ead71d0..72b7d13ee3c6 100644 --- a/fs/nfs/direct.c +++ b/fs/nfs/direct.c @@ -196,6 +196,12 @@ static void nfs_direct_set_hdr_verf(struct nfs_direct_req *dreq, WARN_ON_ONCE(verfp->committed < 0); } +static int nfs_direct_cmp_verf(const struct nfs_writeverf *v1, + const struct nfs_writeverf *v2) +{ + return nfs_write_verifier_cmp(&v1->verifier, &v2->verifier); +} + /* * nfs_direct_cmp_hdr_verf - compare verifier for pgio header * @dreq - direct request possibly spanning multiple servers @@ -215,7 +221,7 @@ static int nfs_direct_set_or_cmp_hdr_verf(struct nfs_direct_req *dreq, nfs_direct_set_hdr_verf(dreq, hdr); return 0; } - return memcmp(verfp, &hdr->verf, sizeof(struct nfs_writeverf)); + return nfs_direct_cmp_verf(verfp, &hdr->verf); } /* @@ -238,7 +244,7 @@ static int nfs_direct_cmp_commit_data_verf(struct nfs_direct_req *dreq, if (verfp->committed < 0) return 1; - return memcmp(verfp, &data->verf, sizeof(struct nfs_writeverf)); + return nfs_direct_cmp_verf(verfp, &data->verf); } /** @@ -366,22 +372,10 @@ out: * Synchronous I/O uses a stack-allocated iocb. Thus we can't trust * the iocb is still valid here if this is a synchronous request. */ -static void nfs_direct_complete(struct nfs_direct_req *dreq, bool write) +static void nfs_direct_complete(struct nfs_direct_req *dreq) { struct inode *inode = dreq->inode; - if (dreq->iocb && write) { - loff_t pos = dreq->iocb->ki_pos + dreq->count; - - spin_lock(&inode->i_lock); - if (i_size_read(inode) < pos) - i_size_write(inode, pos); - spin_unlock(&inode->i_lock); - } - - if (write) - nfs_zap_mapping(inode, inode->i_mapping); - inode_dio_end(inode); if (dreq->iocb) { @@ -436,7 +430,7 @@ static void nfs_direct_read_completion(struct nfs_pgio_header *hdr) } out_put: if (put_dreq(dreq)) - nfs_direct_complete(dreq, false); + nfs_direct_complete(dreq); hdr->release(hdr); } @@ -542,7 +536,7 @@ static ssize_t nfs_direct_read_schedule_iovec(struct nfs_direct_req *dreq, } if (put_dreq(dreq)) - nfs_direct_complete(dreq, false); + nfs_direct_complete(dreq); return 0; } @@ -583,17 +577,12 @@ ssize_t nfs_file_direct_read(struct kiocb *iocb, struct iov_iter *iter) if (!count) goto out; - inode_lock(inode); - result = nfs_sync_mapping(mapping); - if (result) - goto out_unlock; - task_io_account_read(count); result = -ENOMEM; dreq = nfs_direct_req_alloc(); if (dreq == NULL) - goto out_unlock; + goto out; dreq->inode = inode; dreq->bytes_left = dreq->max_count = count; @@ -608,10 +597,12 @@ ssize_t nfs_file_direct_read(struct kiocb *iocb, struct iov_iter *iter) if (!is_sync_kiocb(iocb)) dreq->iocb = iocb; + nfs_start_io_direct(inode); + NFS_I(inode)->read_io += count; result = nfs_direct_read_schedule_iovec(dreq, iter, iocb->ki_pos); - inode_unlock(inode); + nfs_end_io_direct(inode); if (!result) { result = nfs_direct_wait(dreq); @@ -619,13 +610,8 @@ ssize_t nfs_file_direct_read(struct kiocb *iocb, struct iov_iter *iter) iocb->ki_pos += result; } - nfs_direct_req_release(dreq); - return result; - out_release: nfs_direct_req_release(dreq); -out_unlock: - inode_unlock(inode); out: return result; } @@ -657,6 +643,8 @@ static void nfs_direct_write_reschedule(struct nfs_direct_req *dreq) nfs_direct_write_scan_commit_list(dreq->inode, &reqs, &cinfo); dreq->count = 0; + dreq->verf.committed = NFS_INVALID_STABLE_HOW; + nfs_clear_pnfs_ds_commit_verifiers(&dreq->ds_cinfo); for (i = 0; i < dreq->mirror_count; i++) dreq->mirrors[i].count = 0; get_dreq(dreq); @@ -775,7 +763,8 @@ static void nfs_direct_write_schedule_work(struct work_struct *work) nfs_direct_write_reschedule(dreq); break; default: - nfs_direct_complete(dreq, true); + nfs_zap_mapping(dreq->inode, dreq->inode->i_mapping); + nfs_direct_complete(dreq); } } @@ -991,6 +980,7 @@ static ssize_t nfs_direct_write_schedule_iovec(struct nfs_direct_req *dreq, ssize_t nfs_file_direct_write(struct kiocb *iocb, struct iov_iter *iter) { ssize_t result = -EINVAL; + size_t count; struct file *file = iocb->ki_filp; struct address_space *mapping = file->f_mapping; struct inode *inode = mapping->host; @@ -1001,34 +991,24 @@ ssize_t nfs_file_direct_write(struct kiocb *iocb, struct iov_iter *iter) dfprintk(FILE, "NFS: direct write(%pD2, %zd@%Ld)\n", file, iov_iter_count(iter), (long long) iocb->ki_pos); - nfs_add_stats(mapping->host, NFSIOS_DIRECTWRITTENBYTES, - iov_iter_count(iter)); + result = generic_write_checks(iocb, iter); + if (result <= 0) + return result; + count = result; + nfs_add_stats(mapping->host, NFSIOS_DIRECTWRITTENBYTES, count); pos = iocb->ki_pos; end = (pos + iov_iter_count(iter) - 1) >> PAGE_SHIFT; - inode_lock(inode); - - result = nfs_sync_mapping(mapping); - if (result) - goto out_unlock; - - if (mapping->nrpages) { - result = invalidate_inode_pages2_range(mapping, - pos >> PAGE_SHIFT, end); - if (result) - goto out_unlock; - } - - task_io_account_write(iov_iter_count(iter)); + task_io_account_write(count); result = -ENOMEM; dreq = nfs_direct_req_alloc(); if (!dreq) - goto out_unlock; + goto out; dreq->inode = inode; - dreq->bytes_left = dreq->max_count = iov_iter_count(iter); + dreq->bytes_left = dreq->max_count = count; dreq->io_start = pos; dreq->ctx = get_nfs_open_context(nfs_file_open_context(iocb->ki_filp)); l_ctx = nfs_get_lock_context(dreq->ctx); @@ -1040,6 +1020,8 @@ ssize_t nfs_file_direct_write(struct kiocb *iocb, struct iov_iter *iter) if (!is_sync_kiocb(iocb)) dreq->iocb = iocb; + nfs_start_io_direct(inode); + result = nfs_direct_write_schedule_iovec(dreq, iter, pos); if (mapping->nrpages) { @@ -1047,30 +1029,19 @@ ssize_t nfs_file_direct_write(struct kiocb *iocb, struct iov_iter *iter) pos >> PAGE_SHIFT, end); } - inode_unlock(inode); + nfs_end_io_direct(inode); if (!result) { result = nfs_direct_wait(dreq); if (result > 0) { - struct inode *inode = mapping->host; - iocb->ki_pos = pos + result; - spin_lock(&inode->i_lock); - if (i_size_read(inode) < iocb->ki_pos) - i_size_write(inode, iocb->ki_pos); - spin_unlock(&inode->i_lock); - /* XXX: should check the generic_write_sync retval */ generic_write_sync(iocb, result); } } - nfs_direct_req_release(dreq); - return result; - out_release: nfs_direct_req_release(dreq); -out_unlock: - inode_unlock(inode); +out: return result; } diff --git a/fs/nfs/file.c b/fs/nfs/file.c index 717a8d6af52d..7d620970f2e1 100644 --- a/fs/nfs/file.c +++ b/fs/nfs/file.c @@ -170,12 +170,14 @@ nfs_file_read(struct kiocb *iocb, struct iov_iter *to) iocb->ki_filp, iov_iter_count(to), (unsigned long) iocb->ki_pos); - result = nfs_revalidate_mapping_protected(inode, iocb->ki_filp->f_mapping); + nfs_start_io_read(inode); + result = nfs_revalidate_mapping(inode, iocb->ki_filp->f_mapping); if (!result) { result = generic_file_read_iter(iocb, to); if (result > 0) nfs_add_stats(inode, NFSIOS_NORMALREADBYTES, result); } + nfs_end_io_read(inode); return result; } EXPORT_SYMBOL_GPL(nfs_file_read); @@ -191,12 +193,14 @@ nfs_file_splice_read(struct file *filp, loff_t *ppos, dprintk("NFS: splice_read(%pD2, %lu@%Lu)\n", filp, (unsigned long) count, (unsigned long long) *ppos); - res = nfs_revalidate_mapping_protected(inode, filp->f_mapping); + nfs_start_io_read(inode); + res = nfs_revalidate_mapping(inode, filp->f_mapping); if (!res) { res = generic_file_splice_read(filp, ppos, pipe, count, flags); if (res > 0) nfs_add_stats(inode, NFSIOS_NORMALREADBYTES, res); } + nfs_end_io_read(inode); return res; } EXPORT_SYMBOL_GPL(nfs_file_splice_read); @@ -272,16 +276,13 @@ nfs_file_fsync(struct file *file, loff_t start, loff_t end, int datasync) trace_nfs_fsync_enter(inode); - inode_dio_wait(inode); do { ret = filemap_write_and_wait_range(inode->i_mapping, start, end); if (ret != 0) break; - inode_lock(inode); ret = nfs_file_fsync_commit(file, start, end, datasync); if (!ret) ret = pnfs_sync_inode(inode, !!datasync); - inode_unlock(inode); /* * If nfs_file_fsync_commit detected a server reboot, then * resend all dirty pages that might have been covered by @@ -359,19 +360,6 @@ static int nfs_write_begin(struct file *file, struct address_space *mapping, file, mapping->host->i_ino, len, (long long) pos); start: - /* - * Prevent starvation issues if someone is doing a consistency - * sync-to-disk - */ - ret = wait_on_bit_action(&NFS_I(mapping->host)->flags, NFS_INO_FLUSHING, - nfs_wait_bit_killable, TASK_KILLABLE); - if (ret) - return ret; - /* - * Wait for O_DIRECT to complete - */ - inode_dio_wait(mapping->host); - page = grab_cache_page_write_begin(mapping, index, flags); if (!page) return -ENOMEM; @@ -432,7 +420,7 @@ static int nfs_write_end(struct file *file, struct address_space *mapping, return status; NFS_I(mapping->host)->write_io += copied; - if (nfs_ctx_key_to_expire(ctx)) { + if (nfs_ctx_key_to_expire(ctx, mapping->host)) { status = nfs_wb_all(mapping->host); if (status < 0) return status; @@ -470,31 +458,8 @@ static void nfs_invalidate_page(struct page *page, unsigned int offset, */ static int nfs_release_page(struct page *page, gfp_t gfp) { - struct address_space *mapping = page->mapping; - dfprintk(PAGECACHE, "NFS: release_page(%p)\n", page); - /* Always try to initiate a 'commit' if relevant, but only - * wait for it if the caller allows blocking. Even then, - * only wait 1 second and only if the 'bdi' is not congested. - * Waiting indefinitely can cause deadlocks when the NFS - * server is on this machine, when a new TCP connection is - * needed and in other rare cases. There is no particular - * need to wait extensively here. A short wait has the - * benefit that someone else can worry about the freezer. - */ - if (mapping) { - struct nfs_server *nfss = NFS_SERVER(mapping->host); - nfs_commit_inode(mapping->host, 0); - if (gfpflags_allow_blocking(gfp) && - !bdi_write_congested(&nfss->backing_dev_info)) { - wait_on_page_bit_killable_timeout(page, PG_private, - HZ); - if (PagePrivate(page)) - set_bdi_congested(&nfss->backing_dev_info, - BLK_RW_ASYNC); - } - } /* If PagePrivate() is set, then the page is not freeable */ if (PagePrivate(page)) return 0; @@ -604,6 +569,8 @@ static int nfs_vm_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) filp, filp->f_mapping->host->i_ino, (long long)page_offset(page)); + sb_start_pagefault(inode->i_sb); + /* make sure the cache has finished storing the page */ nfs_fscache_wait_on_page_write(NFS_I(inode), page); @@ -630,6 +597,7 @@ static int nfs_vm_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) out_unlock: unlock_page(page); out: + sb_end_pagefault(inode->i_sb); return ret; } @@ -645,7 +613,7 @@ static int nfs_need_check_write(struct file *filp, struct inode *inode) ctx = nfs_file_open_context(filp); if (test_bit(NFS_CONTEXT_ERROR_WRITE, &ctx->flags) || - nfs_ctx_key_to_expire(ctx)) + nfs_ctx_key_to_expire(ctx, inode)) return 1; return 0; } @@ -656,23 +624,17 @@ ssize_t nfs_file_write(struct kiocb *iocb, struct iov_iter *from) struct inode *inode = file_inode(file); unsigned long written = 0; ssize_t result; - size_t count = iov_iter_count(from); result = nfs_key_timeout_notify(file, inode); if (result) return result; - if (iocb->ki_flags & IOCB_DIRECT) { - result = generic_write_checks(iocb, from); - if (result <= 0) - return result; + if (iocb->ki_flags & IOCB_DIRECT) return nfs_file_direct_write(iocb, from); - } dprintk("NFS: write(%pD2, %zu@%Ld)\n", - file, count, (long long) iocb->ki_pos); + file, iov_iter_count(from), (long long) iocb->ki_pos); - result = -EBUSY; if (IS_SWAPFILE(inode)) goto out_swapfile; /* @@ -684,28 +646,33 @@ ssize_t nfs_file_write(struct kiocb *iocb, struct iov_iter *from) goto out; } - result = count; - if (!count) + nfs_start_io_write(inode); + result = generic_write_checks(iocb, from); + if (result > 0) { + current->backing_dev_info = inode_to_bdi(inode); + result = generic_perform_write(file, from, iocb->ki_pos); + current->backing_dev_info = NULL; + } + nfs_end_io_write(inode); + if (result <= 0) goto out; - result = generic_file_write_iter(iocb, from); - if (result > 0) - written = result; + written = generic_write_sync(iocb, result); + iocb->ki_pos += written; /* Return error values */ - if (result >= 0 && nfs_need_check_write(file, inode)) { + if (nfs_need_check_write(file, inode)) { int err = vfs_fsync(file, 0); if (err < 0) result = err; } - if (result > 0) - nfs_add_stats(inode, NFSIOS_NORMALWRITTENBYTES, written); + nfs_add_stats(inode, NFSIOS_NORMALWRITTENBYTES, written); out: return result; out_swapfile: printk(KERN_INFO "NFS: attempt to write to active swap file!\n"); - goto out; + return -EBUSY; } EXPORT_SYMBOL_GPL(nfs_file_write); @@ -780,11 +747,6 @@ do_unlk(struct file *filp, int cmd, struct file_lock *fl, int is_local) } static int -is_time_granular(struct timespec *ts) { - return ((ts->tv_sec == 0) && (ts->tv_nsec <= 1000)); -} - -static int do_setlk(struct file *filp, int cmd, struct file_lock *fl, int is_local) { struct inode *inode = filp->f_mapping->host; @@ -817,12 +779,8 @@ do_setlk(struct file *filp, int cmd, struct file_lock *fl, int is_local) * This makes locking act as a cache coherency point. */ nfs_sync_mapping(filp->f_mapping); - if (!NFS_PROTO(inode)->have_delegation(inode, FMODE_READ)) { - if (is_time_granular(&NFS_SERVER(inode)->time_delta)) - __nfs_revalidate_inode(NFS_SERVER(inode), inode); - else - nfs_zap_caches(inode); - } + if (!NFS_PROTO(inode)->have_delegation(inode, FMODE_READ)) + nfs_zap_mapping(inode, filp->f_mapping); out: return status; } diff --git a/fs/nfs/filelayout/filelayout.c b/fs/nfs/filelayout/filelayout.c index aa59757389dc..a3fc48ba4931 100644 --- a/fs/nfs/filelayout/filelayout.c +++ b/fs/nfs/filelayout/filelayout.c @@ -255,13 +255,16 @@ static int filelayout_read_done_cb(struct rpc_task *task, static void filelayout_set_layoutcommit(struct nfs_pgio_header *hdr) { + loff_t end_offs = 0; if (FILELAYOUT_LSEG(hdr->lseg)->commit_through_mds || - hdr->res.verf->committed != NFS_DATA_SYNC) + hdr->res.verf->committed == NFS_FILE_SYNC) return; + if (hdr->res.verf->committed == NFS_DATA_SYNC) + end_offs = hdr->mds_offset + (loff_t)hdr->res.count; - pnfs_set_layoutcommit(hdr->inode, hdr->lseg, - hdr->mds_offset + hdr->res.count); + /* Note: if the write is unstable, don't set end_offs until commit */ + pnfs_set_layoutcommit(hdr->inode, hdr->lseg, end_offs); dprintk("%s inode %lu pls_end_pos %lu\n", __func__, hdr->inode->i_ino, (unsigned long) NFS_I(hdr->inode)->layout->plh_lwb); } @@ -354,6 +357,12 @@ static int filelayout_write_done_cb(struct rpc_task *task, } filelayout_set_layoutcommit(hdr); + + /* zero out the fattr */ + hdr->fattr.valid = 0; + if (task->tk_status >= 0) + nfs_writeback_update_inode(hdr); + return 0; } @@ -375,8 +384,7 @@ static int filelayout_commit_done_cb(struct rpc_task *task, return -EAGAIN; } - if (data->verf.committed == NFS_UNSTABLE) - pnfs_set_layoutcommit(data->inode, data->lseg, data->lwb); + pnfs_set_layoutcommit(data->inode, data->lseg, data->lwb); return 0; } diff --git a/fs/nfs/flexfilelayout/flexfilelayout.c b/fs/nfs/flexfilelayout/flexfilelayout.c index 0e8018bc9880..e6206eaf2bdf 100644 --- a/fs/nfs/flexfilelayout/flexfilelayout.c +++ b/fs/nfs/flexfilelayout/flexfilelayout.c @@ -1325,15 +1325,16 @@ ff_layout_need_layoutcommit(struct pnfs_layout_segment *lseg) * we always send layoutcommit after DS writes. */ static void -ff_layout_set_layoutcommit(struct nfs_pgio_header *hdr) +ff_layout_set_layoutcommit(struct inode *inode, + struct pnfs_layout_segment *lseg, + loff_t end_offset) { - if (!ff_layout_need_layoutcommit(hdr->lseg)) + if (!ff_layout_need_layoutcommit(lseg)) return; - pnfs_set_layoutcommit(hdr->inode, hdr->lseg, - hdr->mds_offset + hdr->res.count); - dprintk("%s inode %lu pls_end_pos %lu\n", __func__, hdr->inode->i_ino, - (unsigned long) NFS_I(hdr->inode)->layout->plh_lwb); + pnfs_set_layoutcommit(inode, lseg, end_offset); + dprintk("%s inode %lu pls_end_pos %llu\n", __func__, inode->i_ino, + (unsigned long long) NFS_I(inode)->layout->plh_lwb); } static bool @@ -1469,6 +1470,7 @@ static void ff_layout_read_release(void *data) static int ff_layout_write_done_cb(struct rpc_task *task, struct nfs_pgio_header *hdr) { + loff_t end_offs = 0; int err; trace_nfs4_pnfs_write(hdr, task->tk_status); @@ -1494,7 +1496,10 @@ static int ff_layout_write_done_cb(struct rpc_task *task, if (hdr->res.verf->committed == NFS_FILE_SYNC || hdr->res.verf->committed == NFS_DATA_SYNC) - ff_layout_set_layoutcommit(hdr); + end_offs = hdr->mds_offset + (loff_t)hdr->res.count; + + /* Note: if the write is unstable, don't set end_offs until commit */ + ff_layout_set_layoutcommit(hdr->inode, hdr->lseg, end_offs); /* zero out fattr since we don't care DS attr at all */ hdr->fattr.valid = 0; @@ -1530,9 +1535,7 @@ static int ff_layout_commit_done_cb(struct rpc_task *task, return -EAGAIN; } - if (data->verf.committed == NFS_UNSTABLE - && ff_layout_need_layoutcommit(data->lseg)) - pnfs_set_layoutcommit(data->inode, data->lseg, data->lwb); + ff_layout_set_layoutcommit(data->inode, data->lseg, data->lwb); return 0; } diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index dda689d7a8a7..bf4ec5ecc97e 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -662,9 +662,7 @@ int nfs_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat) trace_nfs_getattr_enter(inode); /* Flush out writes to the server in order to update c/mtime. */ if (S_ISREG(inode->i_mode)) { - inode_lock(inode); - err = nfs_sync_inode(inode); - inode_unlock(inode); + err = filemap_write_and_wait(inode->i_mapping); if (err) goto out; } @@ -879,7 +877,10 @@ void nfs_inode_attach_open_context(struct nfs_open_context *ctx) struct nfs_inode *nfsi = NFS_I(inode); spin_lock(&inode->i_lock); - list_add(&ctx->list, &nfsi->open_files); + if (ctx->mode & FMODE_WRITE) + list_add(&ctx->list, &nfsi->open_files); + else + list_add_tail(&ctx->list, &nfsi->open_files); spin_unlock(&inode->i_lock); } EXPORT_SYMBOL_GPL(nfs_inode_attach_open_context); @@ -972,6 +973,13 @@ __nfs_revalidate_inode(struct nfs_server *server, struct inode *inode) if (NFS_STALE(inode)) goto out; + /* pNFS: Attributes aren't updated until we layoutcommit */ + if (S_ISREG(inode->i_mode)) { + status = pnfs_sync_inode(inode, false); + if (status) + goto out; + } + status = -ENOMEM; fattr = nfs_alloc_fattr(); if (fattr == NULL) @@ -1122,14 +1130,12 @@ out: } /** - * __nfs_revalidate_mapping - Revalidate the pagecache + * nfs_revalidate_mapping - Revalidate the pagecache * @inode - pointer to host inode * @mapping - pointer to mapping - * @may_lock - take inode->i_mutex? */ -static int __nfs_revalidate_mapping(struct inode *inode, - struct address_space *mapping, - bool may_lock) +int nfs_revalidate_mapping(struct inode *inode, + struct address_space *mapping) { struct nfs_inode *nfsi = NFS_I(inode); unsigned long *bitlock = &nfsi->flags; @@ -1178,12 +1184,7 @@ static int __nfs_revalidate_mapping(struct inode *inode, nfsi->cache_validity &= ~NFS_INO_INVALID_DATA; spin_unlock(&inode->i_lock); trace_nfs_invalidate_mapping_enter(inode); - if (may_lock) { - inode_lock(inode); - ret = nfs_invalidate_mapping(inode, mapping); - inode_unlock(inode); - } else - ret = nfs_invalidate_mapping(inode, mapping); + ret = nfs_invalidate_mapping(inode, mapping); trace_nfs_invalidate_mapping_exit(inode, ret); clear_bit_unlock(NFS_INO_INVALIDATING, bitlock); @@ -1193,27 +1194,28 @@ out: return ret; } -/** - * nfs_revalidate_mapping - Revalidate the pagecache - * @inode - pointer to host inode - * @mapping - pointer to mapping - */ -int nfs_revalidate_mapping(struct inode *inode, struct address_space *mapping) +static bool nfs_file_has_writers(struct nfs_inode *nfsi) { - return __nfs_revalidate_mapping(inode, mapping, false); + struct inode *inode = &nfsi->vfs_inode; + + assert_spin_locked(&inode->i_lock); + + if (!S_ISREG(inode->i_mode)) + return false; + if (list_empty(&nfsi->open_files)) + return false; + /* Note: This relies on nfsi->open_files being ordered with writers + * being placed at the head of the list. + * See nfs_inode_attach_open_context() + */ + return (list_first_entry(&nfsi->open_files, + struct nfs_open_context, + list)->mode & FMODE_WRITE) == FMODE_WRITE; } -/** - * nfs_revalidate_mapping_protected - Revalidate the pagecache - * @inode - pointer to host inode - * @mapping - pointer to mapping - * - * Differs from nfs_revalidate_mapping() in that it grabs the inode->i_mutex - * while invalidating the mapping. - */ -int nfs_revalidate_mapping_protected(struct inode *inode, struct address_space *mapping) +static bool nfs_file_has_buffered_writers(struct nfs_inode *nfsi) { - return __nfs_revalidate_mapping(inode, mapping, true); + return nfs_file_has_writers(nfsi) && nfs_file_io_is_buffered(nfsi); } static unsigned long nfs_wcc_update_inode(struct inode *inode, struct nfs_fattr *fattr) @@ -1280,22 +1282,24 @@ static int nfs_check_inode_attributes(struct inode *inode, struct nfs_fattr *fat if ((fattr->valid & NFS_ATTR_FATTR_TYPE) && (inode->i_mode & S_IFMT) != (fattr->mode & S_IFMT)) return -EIO; - if ((fattr->valid & NFS_ATTR_FATTR_CHANGE) != 0 && - inode->i_version != fattr->change_attr) - invalid |= NFS_INO_INVALID_ATTR|NFS_INO_REVAL_PAGECACHE; + if (!nfs_file_has_buffered_writers(nfsi)) { + /* Verify a few of the more important attributes */ + if ((fattr->valid & NFS_ATTR_FATTR_CHANGE) != 0 && inode->i_version != fattr->change_attr) + invalid |= NFS_INO_INVALID_ATTR | NFS_INO_REVAL_PAGECACHE; - /* Verify a few of the more important attributes */ - if ((fattr->valid & NFS_ATTR_FATTR_MTIME) && !timespec_equal(&inode->i_mtime, &fattr->mtime)) - invalid |= NFS_INO_INVALID_ATTR; + if ((fattr->valid & NFS_ATTR_FATTR_MTIME) && !timespec_equal(&inode->i_mtime, &fattr->mtime)) + invalid |= NFS_INO_INVALID_ATTR; - if (fattr->valid & NFS_ATTR_FATTR_SIZE) { - cur_size = i_size_read(inode); - new_isize = nfs_size_to_loff_t(fattr->size); - if (cur_size != new_isize) - invalid |= NFS_INO_INVALID_ATTR|NFS_INO_REVAL_PAGECACHE; + if ((fattr->valid & NFS_ATTR_FATTR_CTIME) && !timespec_equal(&inode->i_ctime, &fattr->ctime)) + invalid |= NFS_INO_INVALID_ATTR; + + if (fattr->valid & NFS_ATTR_FATTR_SIZE) { + cur_size = i_size_read(inode); + new_isize = nfs_size_to_loff_t(fattr->size); + if (cur_size != new_isize) + invalid |= NFS_INO_INVALID_ATTR|NFS_INO_REVAL_PAGECACHE; + } } - if (nfsi->nrequests != 0) - invalid &= ~NFS_INO_REVAL_PAGECACHE; /* Have any file permissions changed? */ if ((fattr->valid & NFS_ATTR_FATTR_MODE) && (inode->i_mode & S_IALLUGO) != (fattr->mode & S_IALLUGO)) @@ -1470,28 +1474,12 @@ static int nfs_inode_attrs_need_update(const struct inode *inode, const struct n ((long)nfsi->attr_gencount - (long)nfs_read_attr_generation_counter() > 0); } -/* - * Don't trust the change_attribute, mtime, ctime or size if - * a pnfs LAYOUTCOMMIT is outstanding - */ -static void nfs_inode_attrs_handle_layoutcommit(struct inode *inode, - struct nfs_fattr *fattr) -{ - if (pnfs_layoutcommit_outstanding(inode)) - fattr->valid &= ~(NFS_ATTR_FATTR_CHANGE | - NFS_ATTR_FATTR_MTIME | - NFS_ATTR_FATTR_CTIME | - NFS_ATTR_FATTR_SIZE); -} - static int nfs_refresh_inode_locked(struct inode *inode, struct nfs_fattr *fattr) { int ret; trace_nfs_refresh_inode_enter(inode); - nfs_inode_attrs_handle_layoutcommit(inode, fattr); - if (nfs_inode_attrs_need_update(inode, fattr)) ret = nfs_update_inode(inode, fattr); else @@ -1527,7 +1515,7 @@ EXPORT_SYMBOL_GPL(nfs_refresh_inode); static int nfs_post_op_update_inode_locked(struct inode *inode, struct nfs_fattr *fattr) { - unsigned long invalid = NFS_INO_INVALID_ATTR|NFS_INO_REVAL_PAGECACHE; + unsigned long invalid = NFS_INO_INVALID_ATTR; /* * Don't revalidate the pagecache if we hold a delegation, but do @@ -1676,6 +1664,7 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr) unsigned long invalid = 0; unsigned long now = jiffies; unsigned long save_cache_validity; + bool have_writers = nfs_file_has_buffered_writers(nfsi); bool cache_revalidated = true; dfprintk(VFS, "NFS: %s(%s/%lu fh_crc=0x%08x ct=%d info=0x%x)\n", @@ -1725,17 +1714,25 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr) /* Do atomic weak cache consistency updates */ invalid |= nfs_wcc_update_inode(inode, fattr); + if (pnfs_layoutcommit_outstanding(inode)) { + nfsi->cache_validity |= save_cache_validity & NFS_INO_INVALID_ATTR; + cache_revalidated = false; + } + /* More cache consistency checks */ if (fattr->valid & NFS_ATTR_FATTR_CHANGE) { if (inode->i_version != fattr->change_attr) { dprintk("NFS: change_attr change on server for file %s/%ld\n", inode->i_sb->s_id, inode->i_ino); - invalid |= NFS_INO_INVALID_ATTR - | NFS_INO_INVALID_DATA - | NFS_INO_INVALID_ACCESS - | NFS_INO_INVALID_ACL; - if (S_ISDIR(inode->i_mode)) - nfs_force_lookup_revalidate(inode); + /* Could it be a race with writeback? */ + if (!have_writers) { + invalid |= NFS_INO_INVALID_ATTR + | NFS_INO_INVALID_DATA + | NFS_INO_INVALID_ACCESS + | NFS_INO_INVALID_ACL; + if (S_ISDIR(inode->i_mode)) + nfs_force_lookup_revalidate(inode); + } inode->i_version = fattr->change_attr; } } else { @@ -1768,9 +1765,10 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr) if (new_isize != cur_isize) { /* Do we perhaps have any outstanding writes, or has * the file grown beyond our last write? */ - if ((nfsi->nrequests == 0) || new_isize > cur_isize) { + if (nfsi->nrequests == 0 || new_isize > cur_isize) { i_size_write(inode, new_isize); - invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_DATA; + if (!have_writers) + invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_DATA; } dprintk("NFS: isize change on server for file %s/%ld " "(%Ld to %Ld)\n", diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h index 5ea04d87fc65..7ce5e023c3c3 100644 --- a/fs/nfs/internal.h +++ b/fs/nfs/internal.h @@ -66,13 +66,16 @@ struct nfs_clone_mount { struct nfs_client_initdata { unsigned long init_flags; - const char *hostname; - const struct sockaddr *addr; + const char *hostname; /* Hostname of the server */ + const struct sockaddr *addr; /* Address of the server */ + const char *nodename; /* Hostname of the client */ + const char *ip_addr; /* IP address of the client */ size_t addrlen; struct nfs_subversion *nfs_mod; int proto; u32 minorversion; struct net *net; + const struct rpc_timeout *timeparms; }; /* @@ -147,9 +150,8 @@ extern void nfs_umount(const struct nfs_mount_request *info); extern const struct rpc_program nfs_program; extern void nfs_clients_init(struct net *net); extern struct nfs_client *nfs_alloc_client(const struct nfs_client_initdata *); -int nfs_create_rpc_client(struct nfs_client *, const struct rpc_timeout *, rpc_authflavor_t); +int nfs_create_rpc_client(struct nfs_client *, const struct nfs_client_initdata *, rpc_authflavor_t); struct nfs_client *nfs_get_client(const struct nfs_client_initdata *, - const struct rpc_timeout *, const char *, rpc_authflavor_t); int nfs_probe_fsinfo(struct nfs_server *server, struct nfs_fh *, struct nfs_fattr *); void nfs_server_insert_lists(struct nfs_server *); @@ -184,7 +186,7 @@ extern struct nfs_server *nfs_clone_server(struct nfs_server *, rpc_authflavor_t); extern int nfs_wait_client_init_complete(const struct nfs_client *clp); extern void nfs_mark_client_ready(struct nfs_client *clp, int state); -extern struct nfs_client *nfs4_set_ds_client(struct nfs_client* mds_clp, +extern struct nfs_client *nfs4_set_ds_client(struct nfs_server *mds_srv, const struct sockaddr *ds_addr, int ds_addrlen, int ds_proto, unsigned int ds_timeo, @@ -193,7 +195,7 @@ extern struct nfs_client *nfs4_set_ds_client(struct nfs_client* mds_clp, rpc_authflavor_t au_flavor); extern struct rpc_clnt *nfs4_find_or_create_ds_client(struct nfs_client *, struct inode *); -extern struct nfs_client *nfs3_set_ds_client(struct nfs_client *mds_clp, +extern struct nfs_client *nfs3_set_ds_client(struct nfs_server *mds_srv, const struct sockaddr *ds_addr, int ds_addrlen, int ds_proto, unsigned int ds_timeo, unsigned int ds_retrans, rpc_authflavor_t au_flavor); @@ -338,8 +340,7 @@ nfs4_label_copy(struct nfs4_label *dst, struct nfs4_label *src) /* proc.c */ void nfs_close_context(struct nfs_open_context *ctx, int is_sync); extern struct nfs_client *nfs_init_client(struct nfs_client *clp, - const struct rpc_timeout *timeparms, - const char *ip_addr); + const struct nfs_client_initdata *); /* dir.c */ extern void nfs_force_use_readdirplus(struct inode *dir); @@ -411,6 +412,19 @@ extern void __exit unregister_nfs_fs(void); extern bool nfs_sb_active(struct super_block *sb); extern void nfs_sb_deactive(struct super_block *sb); +/* io.c */ +extern void nfs_start_io_read(struct inode *inode); +extern void nfs_end_io_read(struct inode *inode); +extern void nfs_start_io_write(struct inode *inode); +extern void nfs_end_io_write(struct inode *inode); +extern void nfs_start_io_direct(struct inode *inode); +extern void nfs_end_io_direct(struct inode *inode); + +static inline bool nfs_file_io_is_buffered(struct nfs_inode *nfsi) +{ + return test_bit(NFS_INO_ODIRECT, &nfsi->flags) == 0; +} + /* namespace.c */ #define NFS_PATH_CANONICAL 1 extern char *nfs_path(char **p, struct dentry *dentry, @@ -496,9 +510,29 @@ void nfs_init_cinfo(struct nfs_commit_info *cinfo, struct inode *inode, struct nfs_direct_req *dreq); int nfs_key_timeout_notify(struct file *filp, struct inode *inode); -bool nfs_ctx_key_to_expire(struct nfs_open_context *ctx); +bool nfs_ctx_key_to_expire(struct nfs_open_context *ctx, struct inode *inode); void nfs_pageio_stop_mirroring(struct nfs_pageio_descriptor *pgio); +int nfs_filemap_write_and_wait_range(struct address_space *mapping, + loff_t lstart, loff_t lend); + +#ifdef CONFIG_NFS_V4_1 +static inline +void nfs_clear_pnfs_ds_commit_verifiers(struct pnfs_ds_commit_info *cinfo) +{ + int i; + + for (i = 0; i < cinfo->nbuckets; i++) + cinfo->buckets[i].direct_verf.committed = NFS_INVALID_STABLE_HOW; +} +#else +static inline +void nfs_clear_pnfs_ds_commit_verifiers(struct pnfs_ds_commit_info *cinfo) +{ +} +#endif + + #ifdef CONFIG_MIGRATION extern int nfs_migrate_page(struct address_space *, struct page *, struct page *, enum migrate_mode); @@ -506,6 +540,13 @@ extern int nfs_migrate_page(struct address_space *, #define nfs_migrate_page NULL #endif +static inline int +nfs_write_verifier_cmp(const struct nfs_write_verifier *v1, + const struct nfs_write_verifier *v2) +{ + return memcmp(v1->data, v2->data, sizeof(v1->data)); +} + /* unlink.c */ extern struct rpc_task * nfs_async_rename(struct inode *old_dir, struct inode *new_dir, @@ -521,8 +562,7 @@ extern ssize_t nfs_dreq_bytes_left(struct nfs_direct_req *dreq); /* nfs4proc.c */ extern void __nfs4_read_done_cb(struct nfs_pgio_header *); extern struct nfs_client *nfs4_init_client(struct nfs_client *clp, - const struct rpc_timeout *timeparms, - const char *ip_addr); + const struct nfs_client_initdata *); extern int nfs40_walk_client_list(struct nfs_client *clp, struct nfs_client **result, struct rpc_cred *cred); diff --git a/fs/nfs/io.c b/fs/nfs/io.c new file mode 100644 index 000000000000..1fc5d1ce327e --- /dev/null +++ b/fs/nfs/io.c @@ -0,0 +1,147 @@ +/* + * Copyright (c) 2016 Trond Myklebust + * + * I/O and data path helper functionality. + */ + +#include <linux/types.h> +#include <linux/kernel.h> +#include <linux/bitops.h> +#include <linux/rwsem.h> +#include <linux/fs.h> +#include <linux/nfs_fs.h> + +#include "internal.h" + +/* Call with exclusively locked inode->i_rwsem */ +static void nfs_block_o_direct(struct nfs_inode *nfsi, struct inode *inode) +{ + if (test_bit(NFS_INO_ODIRECT, &nfsi->flags)) { + clear_bit(NFS_INO_ODIRECT, &nfsi->flags); + inode_dio_wait(inode); + } +} + +/** + * nfs_start_io_read - declare the file is being used for buffered reads + * @inode - file inode + * + * Declare that a buffered read operation is about to start, and ensure + * that we block all direct I/O. + * On exit, the function ensures that the NFS_INO_ODIRECT flag is unset, + * and holds a shared lock on inode->i_rwsem to ensure that the flag + * cannot be changed. + * In practice, this means that buffered read operations are allowed to + * execute in parallel, thanks to the shared lock, whereas direct I/O + * operations need to wait to grab an exclusive lock in order to set + * NFS_INO_ODIRECT. + * Note that buffered writes and truncates both take a write lock on + * inode->i_rwsem, meaning that those are serialised w.r.t. the reads. + */ +void +nfs_start_io_read(struct inode *inode) +{ + struct nfs_inode *nfsi = NFS_I(inode); + /* Be an optimist! */ + down_read(&inode->i_rwsem); + if (test_bit(NFS_INO_ODIRECT, &nfsi->flags) == 0) + return; + up_read(&inode->i_rwsem); + /* Slow path.... */ + down_write(&inode->i_rwsem); + nfs_block_o_direct(nfsi, inode); + downgrade_write(&inode->i_rwsem); +} + +/** + * nfs_end_io_read - declare that the buffered read operation is done + * @inode - file inode + * + * Declare that a buffered read operation is done, and release the shared + * lock on inode->i_rwsem. + */ +void +nfs_end_io_read(struct inode *inode) +{ + up_read(&inode->i_rwsem); +} + +/** + * nfs_start_io_write - declare the file is being used for buffered writes + * @inode - file inode + * + * Declare that a buffered read operation is about to start, and ensure + * that we block all direct I/O. + */ +void +nfs_start_io_write(struct inode *inode) +{ + down_write(&inode->i_rwsem); + nfs_block_o_direct(NFS_I(inode), inode); +} + +/** + * nfs_end_io_write - declare that the buffered write operation is done + * @inode - file inode + * + * Declare that a buffered write operation is done, and release the + * lock on inode->i_rwsem. + */ +void +nfs_end_io_write(struct inode *inode) +{ + up_write(&inode->i_rwsem); +} + +/* Call with exclusively locked inode->i_rwsem */ +static void nfs_block_buffered(struct nfs_inode *nfsi, struct inode *inode) +{ + if (!test_bit(NFS_INO_ODIRECT, &nfsi->flags)) { + set_bit(NFS_INO_ODIRECT, &nfsi->flags); + nfs_wb_all(inode); + } +} + +/** + * nfs_end_io_direct - declare the file is being used for direct i/o + * @inode - file inode + * + * Declare that a direct I/O operation is about to start, and ensure + * that we block all buffered I/O. + * On exit, the function ensures that the NFS_INO_ODIRECT flag is set, + * and holds a shared lock on inode->i_rwsem to ensure that the flag + * cannot be changed. + * In practice, this means that direct I/O operations are allowed to + * execute in parallel, thanks to the shared lock, whereas buffered I/O + * operations need to wait to grab an exclusive lock in order to clear + * NFS_INO_ODIRECT. + * Note that buffered writes and truncates both take a write lock on + * inode->i_rwsem, meaning that those are serialised w.r.t. O_DIRECT. + */ +void +nfs_start_io_direct(struct inode *inode) +{ + struct nfs_inode *nfsi = NFS_I(inode); + /* Be an optimist! */ + down_read(&inode->i_rwsem); + if (test_bit(NFS_INO_ODIRECT, &nfsi->flags) != 0) + return; + up_read(&inode->i_rwsem); + /* Slow path.... */ + down_write(&inode->i_rwsem); + nfs_block_buffered(nfsi, inode); + downgrade_write(&inode->i_rwsem); +} + +/** + * nfs_end_io_direct - declare that the direct i/o operation is done + * @inode - file inode + * + * Declare that a direct I/O operation is done, and release the shared + * lock on inode->i_rwsem. + */ +void +nfs_end_io_direct(struct inode *inode) +{ + up_read(&inode->i_rwsem); +} diff --git a/fs/nfs/nfs3client.c b/fs/nfs/nfs3client.c index 9e9fa347a948..ee753547fb0a 100644 --- a/fs/nfs/nfs3client.c +++ b/fs/nfs/nfs3client.c @@ -76,19 +76,23 @@ struct nfs_server *nfs3_clone_server(struct nfs_server *source, * low timeout interval so that if a connection is lost, we retry through * the MDS. */ -struct nfs_client *nfs3_set_ds_client(struct nfs_client *mds_clp, +struct nfs_client *nfs3_set_ds_client(struct nfs_server *mds_srv, const struct sockaddr *ds_addr, int ds_addrlen, int ds_proto, unsigned int ds_timeo, unsigned int ds_retrans, rpc_authflavor_t au_flavor) { + struct rpc_timeout ds_timeout; + struct nfs_client *mds_clp = mds_srv->nfs_client; struct nfs_client_initdata cl_init = { .addr = ds_addr, .addrlen = ds_addrlen, + .nodename = mds_clp->cl_rpcclient->cl_nodename, + .ip_addr = mds_clp->cl_ipaddr, .nfs_mod = &nfs_v3, .proto = ds_proto, .net = mds_clp->cl_net, + .timeparms = &ds_timeout, }; - struct rpc_timeout ds_timeout; struct nfs_client *clp; char buf[INET6_ADDRSTRLEN + 1]; @@ -97,10 +101,12 @@ struct nfs_client *nfs3_set_ds_client(struct nfs_client *mds_clp, return ERR_PTR(-EINVAL); cl_init.hostname = buf; + if (mds_srv->flags & NFS_MOUNT_NORESVPORT) + set_bit(NFS_CS_NORESVPORT, &cl_init.init_flags); + /* Use the MDS nfs_client cl_ipaddr. */ nfs_init_timeout_values(&ds_timeout, ds_proto, ds_timeo, ds_retrans); - clp = nfs_get_client(&cl_init, &ds_timeout, mds_clp->cl_ipaddr, - au_flavor); + clp = nfs_get_client(&cl_init, au_flavor); return clp; } diff --git a/fs/nfs/nfs42proc.c b/fs/nfs/nfs42proc.c index aa03ed09ba06..33da841a21bb 100644 --- a/fs/nfs/nfs42proc.c +++ b/fs/nfs/nfs42proc.c @@ -113,15 +113,17 @@ int nfs42_proc_deallocate(struct file *filep, loff_t offset, loff_t len) if (!nfs_server_capable(inode, NFS_CAP_DEALLOCATE)) return -EOPNOTSUPP; - nfs_wb_all(inode); inode_lock(inode); + err = nfs_sync_inode(inode); + if (err) + goto out_unlock; err = nfs42_proc_fallocate(&msg, filep, offset, len); if (err == 0) truncate_pagecache_range(inode, offset, (offset + len) -1); if (err == -EOPNOTSUPP) NFS_SERVER(inode)->caps &= ~NFS_CAP_DEALLOCATE; - +out_unlock: inode_unlock(inode); return err; } @@ -154,11 +156,20 @@ static ssize_t _nfs42_proc_copy(struct file *src, loff_t pos_src, if (status) return status; + status = nfs_filemap_write_and_wait_range(file_inode(src)->i_mapping, + pos_src, pos_src + (loff_t)count - 1); + if (status) + return status; + status = nfs4_set_rw_stateid(&args.dst_stateid, dst_lock->open_context, dst_lock, FMODE_WRITE); if (status) return status; + status = nfs_sync_inode(dst_inode); + if (status) + return status; + status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0); if (status == -ENOTSUPP) @@ -258,7 +269,11 @@ static loff_t _nfs42_proc_llseek(struct file *filep, if (status) return status; - nfs_wb_all(inode); + status = nfs_filemap_write_and_wait_range(inode->i_mapping, + offset, LLONG_MAX); + if (status) + return status; + status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0); if (status == -ENOTSUPP) @@ -336,8 +351,7 @@ nfs42_layoutstat_done(struct rpc_task *task, void *calldata) * Mark the bad layout state as invalid, then retry * with the current stateid. */ - set_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags); - pnfs_mark_matching_lsegs_invalid(lo, &head, NULL, 0); + pnfs_mark_layout_stateid_invalid(lo, &head); spin_unlock(&inode->i_lock); pnfs_free_lseg_list(&head); } else diff --git a/fs/nfs/nfs42xdr.c b/fs/nfs/nfs42xdr.c index 6dc6f2aea0d6..8b2605882a20 100644 --- a/fs/nfs/nfs42xdr.c +++ b/fs/nfs/nfs42xdr.c @@ -330,13 +330,21 @@ static int decode_write_response(struct xdr_stream *xdr, struct nfs42_write_res *res) { __be32 *p; - int stateids; p = xdr_inline_decode(xdr, 4 + 8 + 4); if (unlikely(!p)) goto out_overflow; - stateids = be32_to_cpup(p++); + /* + * We never use asynchronous mode, so warn if a server returns + * a stateid. + */ + if (unlikely(*p != 0)) { + pr_err_once("%s: server has set unrequested " + "asynchronous mode\n", __func__); + return -EREMOTEIO; + } + p++; p = xdr_decode_hyper(p, &res->count); res->verifier.committed = be32_to_cpup(p); return decode_verifier(xdr, &res->verifier.verifier); diff --git a/fs/nfs/nfs4_fs.h b/fs/nfs/nfs4_fs.h index 768456fa1b17..4be567a54958 100644 --- a/fs/nfs/nfs4_fs.h +++ b/fs/nfs/nfs4_fs.h @@ -185,6 +185,7 @@ struct nfs4_state { struct nfs4_exception { struct nfs4_state *state; struct inode *inode; + nfs4_stateid *stateid; long timeout; unsigned char delay : 1, recovering : 1, diff --git a/fs/nfs/nfs4client.c b/fs/nfs/nfs4client.c index 10410e8b5853..8d7d08d4f95f 100644 --- a/fs/nfs/nfs4client.c +++ b/fs/nfs/nfs4client.c @@ -349,10 +349,10 @@ static int nfs4_init_client_minor_version(struct nfs_client *clp) * Returns pointer to an NFS client, or an ERR_PTR value. */ struct nfs_client *nfs4_init_client(struct nfs_client *clp, - const struct rpc_timeout *timeparms, - const char *ip_addr) + const struct nfs_client_initdata *cl_init) { char buf[INET6_ADDRSTRLEN + 1]; + const char *ip_addr = cl_init->ip_addr; struct nfs_client *old; int error; @@ -370,9 +370,9 @@ struct nfs_client *nfs4_init_client(struct nfs_client *clp, __set_bit(NFS_CS_DISCRTRY, &clp->cl_flags); __set_bit(NFS_CS_NO_RETRANS_TIMEOUT, &clp->cl_flags); - error = nfs_create_rpc_client(clp, timeparms, RPC_AUTH_GSS_KRB5I); + error = nfs_create_rpc_client(clp, cl_init, RPC_AUTH_GSS_KRB5I); if (error == -EINVAL) - error = nfs_create_rpc_client(clp, timeparms, RPC_AUTH_UNIX); + error = nfs_create_rpc_client(clp, cl_init, RPC_AUTH_UNIX); if (error < 0) goto error; @@ -793,10 +793,12 @@ static int nfs4_set_client(struct nfs_server *server, .hostname = hostname, .addr = addr, .addrlen = addrlen, + .ip_addr = ip_addr, .nfs_mod = &nfs_v4, .proto = proto, .minorversion = minorversion, .net = net, + .timeparms = timeparms, }; struct nfs_client *clp; int error; @@ -809,7 +811,7 @@ static int nfs4_set_client(struct nfs_server *server, set_bit(NFS_CS_MIGRATION, &cl_init.init_flags); /* Allocate or find a client reference we can use */ - clp = nfs_get_client(&cl_init, timeparms, ip_addr, authflavour); + clp = nfs_get_client(&cl_init, authflavour); if (IS_ERR(clp)) { error = PTR_ERR(clp); goto error; @@ -842,20 +844,24 @@ error: * low timeout interval so that if a connection is lost, we retry through * the MDS. */ -struct nfs_client *nfs4_set_ds_client(struct nfs_client* mds_clp, +struct nfs_client *nfs4_set_ds_client(struct nfs_server *mds_srv, const struct sockaddr *ds_addr, int ds_addrlen, int ds_proto, unsigned int ds_timeo, unsigned int ds_retrans, u32 minor_version, rpc_authflavor_t au_flavor) { + struct rpc_timeout ds_timeout; + struct nfs_client *mds_clp = mds_srv->nfs_client; struct nfs_client_initdata cl_init = { .addr = ds_addr, .addrlen = ds_addrlen, + .nodename = mds_clp->cl_rpcclient->cl_nodename, + .ip_addr = mds_clp->cl_ipaddr, .nfs_mod = &nfs_v4, .proto = ds_proto, .minorversion = minor_version, .net = mds_clp->cl_net, + .timeparms = &ds_timeout, }; - struct rpc_timeout ds_timeout; struct nfs_client *clp; char buf[INET6_ADDRSTRLEN + 1]; @@ -863,14 +869,16 @@ struct nfs_client *nfs4_set_ds_client(struct nfs_client* mds_clp, return ERR_PTR(-EINVAL); cl_init.hostname = buf; + if (mds_srv->flags & NFS_MOUNT_NORESVPORT) + __set_bit(NFS_CS_NORESVPORT, &cl_init.init_flags); + /* * Set an authflavor equual to the MDS value. Use the MDS nfs_client * cl_ipaddr so as to use the same EXCHANGE_ID co_ownerid as the MDS * (section 13.1 RFC 5661). */ nfs_init_timeout_values(&ds_timeout, ds_proto, ds_timeo, ds_retrans); - clp = nfs_get_client(&cl_init, &ds_timeout, mds_clp->cl_ipaddr, - au_flavor); + clp = nfs_get_client(&cl_init, au_flavor); dprintk("<-- %s %p\n", __func__, clp); return clp; diff --git a/fs/nfs/nfs4file.c b/fs/nfs/nfs4file.c index 014b0e41ace5..d085ad794884 100644 --- a/fs/nfs/nfs4file.c +++ b/fs/nfs/nfs4file.c @@ -66,7 +66,7 @@ nfs4_file_open(struct inode *inode, struct file *filp) if (openflags & O_TRUNC) { attr.ia_valid |= ATTR_SIZE; attr.ia_size = 0; - nfs_sync_inode(inode); + filemap_write_and_wait(inode->i_mapping); } inode = NFS_PROTO(dir)->open_context(dir, ctx, openflags, &attr, NULL); @@ -133,21 +133,9 @@ static ssize_t nfs4_copy_file_range(struct file *file_in, loff_t pos_in, struct file *file_out, loff_t pos_out, size_t count, unsigned int flags) { - struct inode *in_inode = file_inode(file_in); - struct inode *out_inode = file_inode(file_out); - int ret; - - if (in_inode == out_inode) + if (file_inode(file_in) == file_inode(file_out)) return -EINVAL; - /* flush any pending writes */ - ret = nfs_sync_inode(in_inode); - if (ret) - return ret; - ret = nfs_sync_inode(out_inode); - if (ret) - return ret; - return nfs42_proc_copy(file_in, pos_in, file_out, pos_out, count); } diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index ff416d0e24bc..da5c9e58e907 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -363,6 +363,7 @@ static int nfs4_do_handle_exception(struct nfs_server *server, { struct nfs_client *clp = server->nfs_client; struct nfs4_state *state = exception->state; + const nfs4_stateid *stateid = exception->stateid; struct inode *inode = exception->inode; int ret = errorcode; @@ -376,9 +377,18 @@ static int nfs4_do_handle_exception(struct nfs_server *server, case -NFS4ERR_DELEG_REVOKED: case -NFS4ERR_ADMIN_REVOKED: case -NFS4ERR_BAD_STATEID: - if (inode && nfs_async_inode_return_delegation(inode, - NULL) == 0) - goto wait_on_recovery; + if (inode) { + int err; + + err = nfs_async_inode_return_delegation(inode, + stateid); + if (err == 0) + goto wait_on_recovery; + if (stateid != NULL && stateid->type == NFS4_DELEGATION_STATEID_TYPE) { + exception->retry = 1; + break; + } + } if (state == NULL) break; ret = nfs4_schedule_stateid_recovery(server, state); @@ -427,6 +437,7 @@ static int nfs4_do_handle_exception(struct nfs_server *server, case -NFS4ERR_DELAY: nfs_inc_server_stats(server, NFSIOS_DELAY); case -NFS4ERR_GRACE: + case -NFS4ERR_LAYOUTTRYLATER: case -NFS4ERR_RECALLCONFLICT: exception->delay = 1; return 0; @@ -2669,28 +2680,17 @@ static struct nfs4_state *nfs4_do_open(struct inode *dir, return res; } -static int _nfs4_do_setattr(struct inode *inode, struct rpc_cred *cred, - struct nfs_fattr *fattr, struct iattr *sattr, - struct nfs4_state *state, struct nfs4_label *ilabel, - struct nfs4_label *olabel) +static int _nfs4_do_setattr(struct inode *inode, + struct nfs_setattrargs *arg, + struct nfs_setattrres *res, + struct rpc_cred *cred, + struct nfs4_state *state) { struct nfs_server *server = NFS_SERVER(inode); - struct nfs_setattrargs arg = { - .fh = NFS_FH(inode), - .iap = sattr, - .server = server, - .bitmask = server->attr_bitmask, - .label = ilabel, - }; - struct nfs_setattrres res = { - .fattr = fattr, - .label = olabel, - .server = server, - }; struct rpc_message msg = { .rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SETATTR], - .rpc_argp = &arg, - .rpc_resp = &res, + .rpc_argp = arg, + .rpc_resp = res, .rpc_cred = cred, }; struct rpc_cred *delegation_cred = NULL; @@ -2699,17 +2699,13 @@ static int _nfs4_do_setattr(struct inode *inode, struct rpc_cred *cred, bool truncate; int status; - arg.bitmask = nfs4_bitmask(server, ilabel); - if (ilabel) - arg.bitmask = nfs4_bitmask(server, olabel); - - nfs_fattr_init(fattr); + nfs_fattr_init(res->fattr); /* Servers should only apply open mode checks for file size changes */ - truncate = (sattr->ia_valid & ATTR_SIZE) ? true : false; + truncate = (arg->iap->ia_valid & ATTR_SIZE) ? true : false; fmode = truncate ? FMODE_WRITE : FMODE_READ; - if (nfs4_copy_delegation_stateid(inode, fmode, &arg.stateid, &delegation_cred)) { + if (nfs4_copy_delegation_stateid(inode, fmode, &arg->stateid, &delegation_cred)) { /* Use that stateid */ } else if (truncate && state != NULL) { struct nfs_lockowner lockowner = { @@ -2719,19 +2715,19 @@ static int _nfs4_do_setattr(struct inode *inode, struct rpc_cred *cred, if (!nfs4_valid_open_stateid(state)) return -EBADF; if (nfs4_select_rw_stateid(state, FMODE_WRITE, &lockowner, - &arg.stateid, &delegation_cred) == -EIO) + &arg->stateid, &delegation_cred) == -EIO) return -EBADF; } else - nfs4_stateid_copy(&arg.stateid, &zero_stateid); + nfs4_stateid_copy(&arg->stateid, &zero_stateid); if (delegation_cred) msg.rpc_cred = delegation_cred; - status = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1); + status = nfs4_call_sync(server->client, server, &msg, &arg->seq_args, &res->seq_res, 1); put_rpccred(delegation_cred); if (status == 0 && state != NULL) renew_lease(server, timestamp); - trace_nfs4_setattr(inode, &arg.stateid, status); + trace_nfs4_setattr(inode, &arg->stateid, status); return status; } @@ -2741,13 +2737,31 @@ static int nfs4_do_setattr(struct inode *inode, struct rpc_cred *cred, struct nfs4_label *olabel) { struct nfs_server *server = NFS_SERVER(inode); + struct nfs_setattrargs arg = { + .fh = NFS_FH(inode), + .iap = sattr, + .server = server, + .bitmask = server->attr_bitmask, + .label = ilabel, + }; + struct nfs_setattrres res = { + .fattr = fattr, + .label = olabel, + .server = server, + }; struct nfs4_exception exception = { .state = state, .inode = inode, + .stateid = &arg.stateid, }; int err; + + arg.bitmask = nfs4_bitmask(server, ilabel); + if (ilabel) + arg.bitmask = nfs4_bitmask(server, olabel); + do { - err = _nfs4_do_setattr(inode, cred, fattr, sattr, state, ilabel, olabel); + err = _nfs4_do_setattr(inode, &arg, &res, cred, state); switch (err) { case -NFS4ERR_OPENMODE: if (!(sattr->ia_valid & ATTR_SIZE)) { @@ -3267,13 +3281,6 @@ static int nfs4_find_root_sec(struct nfs_server *server, struct nfs_fh *fhandle, return status; } -static int nfs4_do_find_root_sec(struct nfs_server *server, - struct nfs_fh *fhandle, struct nfs_fsinfo *info) -{ - int mv = server->nfs_client->cl_minorversion; - return nfs_v4_minor_ops[mv]->find_root_sec(server, fhandle, info); -} - /** * nfs4_proc_get_rootfh - get file handle for server's pseudoroot * @server: initialized nfs_server handle @@ -3293,7 +3300,8 @@ int nfs4_proc_get_rootfh(struct nfs_server *server, struct nfs_fh *fhandle, status = nfs4_lookup_root(server, fhandle, info); if (auth_probe || status == NFS4ERR_WRONGSEC) - status = nfs4_do_find_root_sec(server, fhandle, info); + status = server->nfs_client->cl_mvops->find_root_sec(server, + fhandle, info); if (status == 0) status = nfs4_server_capabilities(server, fhandle); @@ -4392,7 +4400,8 @@ static void nfs4_proc_read_setup(struct nfs_pgio_header *hdr, struct rpc_message *msg) { hdr->timestamp = jiffies; - hdr->pgio_done_cb = nfs4_read_done_cb; + if (!hdr->pgio_done_cb) + hdr->pgio_done_cb = nfs4_read_done_cb; msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_READ]; nfs4_init_sequence(&hdr->args.seq_args, &hdr->res.seq_res, 0); } @@ -7869,11 +7878,13 @@ nfs4_layoutget_handle_exception(struct rpc_task *task, struct inode *inode = lgp->args.inode; struct nfs_server *server = NFS_SERVER(inode); struct pnfs_layout_hdr *lo; - int status = task->tk_status; + int nfs4err = task->tk_status; + int err, status = 0; + LIST_HEAD(head); dprintk("--> %s tk_status => %d\n", __func__, -task->tk_status); - switch (status) { + switch (nfs4err) { case 0: goto out; @@ -7905,45 +7916,42 @@ nfs4_layoutget_handle_exception(struct rpc_task *task, status = -EOVERFLOW; goto out; } - /* Fallthrough */ + status = -EBUSY; + break; case -NFS4ERR_RECALLCONFLICT: - nfs4_handle_exception(server, -NFS4ERR_RECALLCONFLICT, - exception); status = -ERECALLCONFLICT; - goto out; + break; case -NFS4ERR_EXPIRED: case -NFS4ERR_BAD_STATEID: exception->timeout = 0; spin_lock(&inode->i_lock); - if (nfs4_stateid_match(&lgp->args.stateid, + lo = NFS_I(inode)->layout; + /* If the open stateid was bad, then recover it. */ + if (!lo || test_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags) || + nfs4_stateid_match_other(&lgp->args.stateid, &lgp->args.ctx->state->stateid)) { spin_unlock(&inode->i_lock); - /* If the open stateid was bad, then recover it. */ exception->state = lgp->args.ctx->state; break; } - lo = NFS_I(inode)->layout; - if (lo && !test_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags) && - nfs4_stateid_match_other(&lgp->args.stateid, &lo->plh_stateid)) { - LIST_HEAD(head); - - /* - * Mark the bad layout state as invalid, then retry - * with the current stateid. - */ - set_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags); - pnfs_mark_matching_lsegs_invalid(lo, &head, NULL, 0); - spin_unlock(&inode->i_lock); - pnfs_free_lseg_list(&head); - status = -EAGAIN; - goto out; - } else - spin_unlock(&inode->i_lock); - } - status = nfs4_handle_exception(server, status, exception); - if (exception->retry) + /* + * Mark the bad layout state as invalid, then retry + */ + pnfs_mark_layout_stateid_invalid(lo, &head); + spin_unlock(&inode->i_lock); + pnfs_free_lseg_list(&head); status = -EAGAIN; + goto out; + } + + err = nfs4_handle_exception(server, nfs4err, exception); + if (!status) { + if (exception->retry) + status = -EAGAIN; + else + status = err; + } out: dprintk("<-- %s\n", __func__); return status; @@ -8129,8 +8137,7 @@ static void nfs4_layoutreturn_release(void *calldata) spin_lock(&lo->plh_inode->i_lock); pnfs_mark_matching_lsegs_invalid(lo, &freeme, &lrp->args.range, be32_to_cpu(lrp->args.stateid.seqid)); - pnfs_mark_layout_returned_if_empty(lo); - if (lrp->res.lrs_present) + if (lrp->res.lrs_present && pnfs_layout_is_valid(lo)) pnfs_set_layout_stateid(lo, &lrp->res.stateid, true); pnfs_clear_layoutreturn_waitbit(lo); spin_unlock(&lo->plh_inode->i_lock); @@ -8835,7 +8842,7 @@ const struct nfs4_minor_version_ops *nfs_v4_minor_ops[] = { #endif }; -ssize_t nfs4_listxattr(struct dentry *dentry, char *list, size_t size) +static ssize_t nfs4_listxattr(struct dentry *dentry, char *list, size_t size) { ssize_t error, error2; diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c index 661e753fe1c9..7bd3a5c09d31 100644 --- a/fs/nfs/nfs4xdr.c +++ b/fs/nfs/nfs4xdr.c @@ -1985,9 +1985,14 @@ encode_layoutcommit(struct xdr_stream *xdr, p = xdr_encode_hyper(p, args->lastbytewritten + 1); /* length */ *p = cpu_to_be32(0); /* reclaim */ encode_nfs4_stateid(xdr, &args->stateid); - p = reserve_space(xdr, 20); - *p++ = cpu_to_be32(1); /* newoffset = TRUE */ - p = xdr_encode_hyper(p, args->lastbytewritten); + if (args->lastbytewritten != U64_MAX) { + p = reserve_space(xdr, 20); + *p++ = cpu_to_be32(1); /* newoffset = TRUE */ + p = xdr_encode_hyper(p, args->lastbytewritten); + } else { + p = reserve_space(xdr, 12); + *p++ = cpu_to_be32(0); /* newoffset = FALSE */ + } *p++ = cpu_to_be32(0); /* Never send time_modify_changed */ *p++ = cpu_to_be32(NFS_SERVER(args->inode)->pnfs_curr_ld->id);/* type */ diff --git a/fs/nfs/nfstrace.h b/fs/nfs/nfstrace.h index 31c7763b94d5..2ca9167bc97d 100644 --- a/fs/nfs/nfstrace.h +++ b/fs/nfs/nfstrace.h @@ -37,7 +37,6 @@ { 1 << NFS_INO_ADVISE_RDPLUS, "ADVISE_RDPLUS" }, \ { 1 << NFS_INO_STALE, "STALE" }, \ { 1 << NFS_INO_INVALIDATING, "INVALIDATING" }, \ - { 1 << NFS_INO_FLUSHING, "FLUSHING" }, \ { 1 << NFS_INO_FSCACHE, "FSCACHE" }, \ { 1 << NFS_INO_LAYOUTCOMMIT, "NEED_LAYOUTCOMMIT" }, \ { 1 << NFS_INO_LAYOUTCOMMITTING, "LAYOUTCOMMIT" }) diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index 0fbe734cc38c..70806cae0d36 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -259,7 +259,7 @@ pnfs_put_layout_hdr(struct pnfs_layout_hdr *lo) * is required. * Note that caller must hold inode->i_lock. */ -static int +int pnfs_mark_layout_stateid_invalid(struct pnfs_layout_hdr *lo, struct list_head *lseg_list) { @@ -334,14 +334,17 @@ pnfs_layout_io_test_failed(struct pnfs_layout_hdr *lo, u32 iomode) } static void -init_lseg(struct pnfs_layout_hdr *lo, struct pnfs_layout_segment *lseg) +pnfs_init_lseg(struct pnfs_layout_hdr *lo, struct pnfs_layout_segment *lseg, + const struct pnfs_layout_range *range, + const nfs4_stateid *stateid) { INIT_LIST_HEAD(&lseg->pls_list); INIT_LIST_HEAD(&lseg->pls_lc_list); atomic_set(&lseg->pls_refcount, 1); - smp_mb(); set_bit(NFS_LSEG_VALID, &lseg->pls_flags); lseg->pls_layout = lo; + lseg->pls_range = *range; + lseg->pls_seq = be32_to_cpu(stateid->seqid); } static void pnfs_free_lseg(struct pnfs_layout_segment *lseg) @@ -486,15 +489,6 @@ pnfs_lseg_range_intersecting(const struct pnfs_layout_range *l1, (end2 == NFS4_MAX_UINT64 || end2 > start1); } -static bool -should_free_lseg(const struct pnfs_layout_range *lseg_range, - const struct pnfs_layout_range *recall_range) -{ - return (recall_range->iomode == IOMODE_ANY || - lseg_range->iomode == recall_range->iomode) && - pnfs_lseg_range_intersecting(lseg_range, recall_range); -} - static bool pnfs_lseg_dec_and_remove_zero(struct pnfs_layout_segment *lseg, struct list_head *tmp_list) { @@ -533,6 +527,27 @@ static bool pnfs_seqid_is_newer(u32 s1, u32 s2) return (s32)(s1 - s2) > 0; } +static bool +pnfs_should_free_range(const struct pnfs_layout_range *lseg_range, + const struct pnfs_layout_range *recall_range) +{ + return (recall_range->iomode == IOMODE_ANY || + lseg_range->iomode == recall_range->iomode) && + pnfs_lseg_range_intersecting(lseg_range, recall_range); +} + +static bool +pnfs_match_lseg_recall(const struct pnfs_layout_segment *lseg, + const struct pnfs_layout_range *recall_range, + u32 seq) +{ + if (seq != 0 && pnfs_seqid_is_newer(lseg->pls_seq, seq)) + return false; + if (recall_range == NULL) + return true; + return pnfs_should_free_range(&lseg->pls_range, recall_range); +} + /** * pnfs_mark_matching_lsegs_invalid - tear down lsegs or mark them for later * @lo: layout header containing the lsegs @@ -562,10 +577,7 @@ pnfs_mark_matching_lsegs_invalid(struct pnfs_layout_hdr *lo, if (list_empty(&lo->plh_segs)) return 0; list_for_each_entry_safe(lseg, next, &lo->plh_segs, pls_list) - if (!recall_range || - should_free_lseg(&lseg->pls_range, recall_range)) { - if (seq && pnfs_seqid_is_newer(lseg->pls_seq, seq)) - continue; + if (pnfs_match_lseg_recall(lseg, recall_range, seq)) { dprintk("%s: freeing lseg %p iomode %d seq %u" "offset %llu length %llu\n", __func__, lseg, lseg->pls_range.iomode, lseg->pls_seq, @@ -761,24 +773,25 @@ void pnfs_set_layout_stateid(struct pnfs_layout_hdr *lo, const nfs4_stateid *new, bool update_barrier) { - u32 oldseq, newseq, new_barrier; - int empty = list_empty(&lo->plh_segs); + u32 oldseq, newseq, new_barrier = 0; + bool invalid = !pnfs_layout_is_valid(lo); oldseq = be32_to_cpu(lo->plh_stateid.seqid); newseq = be32_to_cpu(new->seqid); - if (empty || pnfs_seqid_is_newer(newseq, oldseq)) { + if (invalid || pnfs_seqid_is_newer(newseq, oldseq)) { nfs4_stateid_copy(&lo->plh_stateid, new); - if (update_barrier) { - new_barrier = be32_to_cpu(new->seqid); - } else { - /* Because of wraparound, we want to keep the barrier - * "close" to the current seqids. - */ - new_barrier = newseq - atomic_read(&lo->plh_outstanding); - } - if (empty || pnfs_seqid_is_newer(new_barrier, lo->plh_barrier)) - lo->plh_barrier = new_barrier; + /* + * Because of wraparound, we want to keep the barrier + * "close" to the current seqids. + */ + new_barrier = newseq - atomic_read(&lo->plh_outstanding); } + if (update_barrier) + new_barrier = be32_to_cpu(new->seqid); + else if (new_barrier == 0) + return; + if (invalid || pnfs_seqid_is_newer(new_barrier, lo->plh_barrier)) + lo->plh_barrier = new_barrier; } static bool @@ -873,15 +886,37 @@ void pnfs_clear_layoutreturn_waitbit(struct pnfs_layout_hdr *lo) rpc_wake_up(&NFS_SERVER(lo->plh_inode)->roc_rpcwaitq); } +static void +pnfs_clear_layoutreturn_info(struct pnfs_layout_hdr *lo) +{ + lo->plh_return_iomode = 0; + lo->plh_return_seq = 0; + clear_bit(NFS_LAYOUT_RETURN_REQUESTED, &lo->plh_flags); +} + static bool -pnfs_prepare_layoutreturn(struct pnfs_layout_hdr *lo) +pnfs_prepare_layoutreturn(struct pnfs_layout_hdr *lo, + nfs4_stateid *stateid, + enum pnfs_iomode *iomode) { if (test_and_set_bit(NFS_LAYOUT_RETURN, &lo->plh_flags)) return false; - lo->plh_return_iomode = 0; - lo->plh_return_seq = 0; pnfs_get_layout_hdr(lo); - clear_bit(NFS_LAYOUT_RETURN_REQUESTED, &lo->plh_flags); + if (test_bit(NFS_LAYOUT_RETURN_REQUESTED, &lo->plh_flags)) { + if (stateid != NULL) { + nfs4_stateid_copy(stateid, &lo->plh_stateid); + if (lo->plh_return_seq != 0) + stateid->seqid = cpu_to_be32(lo->plh_return_seq); + } + if (iomode != NULL) + *iomode = lo->plh_return_iomode; + pnfs_clear_layoutreturn_info(lo); + return true; + } + if (stateid != NULL) + nfs4_stateid_copy(stateid, &lo->plh_stateid); + if (iomode != NULL) + *iomode = IOMODE_ANY; return true; } @@ -949,10 +984,7 @@ static void pnfs_layoutreturn_before_put_layout_hdr(struct pnfs_layout_hdr *lo) enum pnfs_iomode iomode; bool send; - nfs4_stateid_copy(&stateid, &lo->plh_stateid); - stateid.seqid = cpu_to_be32(lo->plh_return_seq); - iomode = lo->plh_return_iomode; - send = pnfs_prepare_layoutreturn(lo); + send = pnfs_prepare_layoutreturn(lo, &stateid, &iomode); spin_unlock(&inode->i_lock); if (send) { /* Send an async layoutreturn so we dont deadlock */ @@ -989,7 +1021,6 @@ _pnfs_return_layout(struct inode *ino) dprintk("NFS: %s no layout to return\n", __func__); goto out; } - nfs4_stateid_copy(&stateid, &nfsi->layout->plh_stateid); /* Reference matched in nfs4_layoutreturn_release */ pnfs_get_layout_hdr(lo); empty = list_empty(&lo->plh_segs); @@ -1012,8 +1043,7 @@ _pnfs_return_layout(struct inode *ino) goto out_put_layout_hdr; } - set_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags); - send = pnfs_prepare_layoutreturn(lo); + send = pnfs_prepare_layoutreturn(lo, &stateid, NULL); spin_unlock(&ino->i_lock); pnfs_free_lseg_list(&tmp_list); if (send) @@ -1080,11 +1110,10 @@ bool pnfs_roc(struct inode *ino) goto out_noroc; } - nfs4_stateid_copy(&stateid, &lo->plh_stateid); /* always send layoutreturn if being marked so */ - if (test_and_clear_bit(NFS_LAYOUT_RETURN_REQUESTED, - &lo->plh_flags)) - layoutreturn = pnfs_prepare_layoutreturn(lo); + if (test_bit(NFS_LAYOUT_RETURN_REQUESTED, &lo->plh_flags)) + layoutreturn = pnfs_prepare_layoutreturn(lo, + &stateid, NULL); list_for_each_entry_safe(lseg, tmp, &lo->plh_segs, pls_list) /* If we are sending layoutreturn, invalidate all valid lsegs */ @@ -1132,7 +1161,6 @@ void pnfs_roc_set_barrier(struct inode *ino, u32 barrier) spin_lock(&ino->i_lock); lo = NFS_I(ino)->layout; - pnfs_mark_layout_returned_if_empty(lo); if (pnfs_seqid_is_newer(barrier, lo->plh_barrier)) lo->plh_barrier = barrier; spin_unlock(&ino->i_lock); @@ -1505,7 +1533,7 @@ pnfs_update_layout(struct inode *ino, struct pnfs_layout_segment *lseg = NULL; nfs4_stateid stateid; long timeout = 0; - unsigned long giveup = jiffies + rpc_get_timeout(server->client); + unsigned long giveup = jiffies + (clp->cl_lease_time << 1); bool first; if (!pnfs_enabled_sb(NFS_SERVER(ino))) { @@ -1645,33 +1673,44 @@ lookup_again: lseg = send_layoutget(lo, ctx, &stateid, &arg, &timeout, gfp_flags); trace_pnfs_update_layout(ino, pos, count, iomode, lo, lseg, PNFS_UPDATE_LAYOUT_SEND_LAYOUTGET); + atomic_dec(&lo->plh_outstanding); if (IS_ERR(lseg)) { switch(PTR_ERR(lseg)) { - case -ERECALLCONFLICT: + case -EBUSY: if (time_after(jiffies, giveup)) lseg = NULL; - /* Fallthrough */ - case -EAGAIN: - pnfs_put_layout_hdr(lo); - if (first) - pnfs_clear_first_layoutget(lo); - if (lseg) { - trace_pnfs_update_layout(ino, pos, count, - iomode, lo, lseg, PNFS_UPDATE_LAYOUT_RETRY); - goto lookup_again; + break; + case -ERECALLCONFLICT: + /* Huh? We hold no layouts, how is there a recall? */ + if (first) { + lseg = NULL; + break; } + /* Destroy the existing layout and start over */ + if (time_after(jiffies, giveup)) + pnfs_destroy_layout(NFS_I(ino)); /* Fallthrough */ + case -EAGAIN: + break; default: if (!nfs_error_is_fatal(PTR_ERR(lseg))) { pnfs_layout_clear_fail_bit(lo, pnfs_iomode_to_fail_bit(iomode)); lseg = NULL; } + goto out_put_layout_hdr; + } + if (lseg) { + if (first) + pnfs_clear_first_layoutget(lo); + trace_pnfs_update_layout(ino, pos, count, + iomode, lo, lseg, PNFS_UPDATE_LAYOUT_RETRY); + pnfs_put_layout_hdr(lo); + goto lookup_again; } } else { pnfs_layout_clear_fail_bit(lo, pnfs_iomode_to_fail_bit(iomode)); } - atomic_dec(&lo->plh_outstanding); out_put_layout_hdr: if (first) pnfs_clear_first_layoutget(lo); @@ -1735,9 +1774,7 @@ pnfs_layout_process(struct nfs4_layoutget *lgp) return lseg; } - init_lseg(lo, lseg); - lseg->pls_range = res->range; - lseg->pls_seq = be32_to_cpu(res->stateid.seqid); + pnfs_init_lseg(lo, lseg, &res->range, &res->stateid); spin_lock(&ino->i_lock); if (pnfs_layoutgets_blocked(lo)) { @@ -1758,16 +1795,19 @@ pnfs_layout_process(struct nfs4_layoutget *lgp) * inode invalid, and don't bother validating the stateid * sequence number. */ - pnfs_mark_matching_lsegs_invalid(lo, &free_me, NULL, 0); + pnfs_mark_layout_stateid_invalid(lo, &free_me); nfs4_stateid_copy(&lo->plh_stateid, &res->stateid); lo->plh_barrier = be32_to_cpu(res->stateid.seqid); } - clear_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags); - pnfs_get_lseg(lseg); pnfs_layout_insert_lseg(lo, lseg, &free_me); + if (!pnfs_layout_is_valid(lo)) { + pnfs_clear_layoutreturn_info(lo); + clear_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags); + } + if (res->return_on_close) set_bit(NFS_LSEG_ROC, &lseg->pls_flags); @@ -1787,14 +1827,14 @@ static void pnfs_set_plh_return_info(struct pnfs_layout_hdr *lo, enum pnfs_iomode iomode, u32 seq) { - if (lo->plh_return_iomode == iomode) - return; - if (lo->plh_return_iomode != 0) + if (lo->plh_return_iomode != 0 && lo->plh_return_iomode != iomode) iomode = IOMODE_ANY; lo->plh_return_iomode = iomode; set_bit(NFS_LAYOUT_RETURN_REQUESTED, &lo->plh_flags); - if (!lo->plh_return_seq || pnfs_seqid_is_newer(seq, lo->plh_return_seq)) + if (seq != 0) { + WARN_ON_ONCE(lo->plh_return_seq != 0 && lo->plh_return_seq != seq); lo->plh_return_seq = seq; + } } /** @@ -1824,7 +1864,7 @@ pnfs_mark_matching_lsegs_return(struct pnfs_layout_hdr *lo, assert_spin_locked(&lo->plh_inode->i_lock); list_for_each_entry_safe(lseg, next, &lo->plh_segs, pls_list) - if (should_free_lseg(&lseg->pls_range, return_range)) { + if (pnfs_match_lseg_recall(lseg, return_range, seq)) { dprintk("%s: marking lseg %p iomode %d " "offset %llu length %llu\n", __func__, lseg, lseg->pls_range.iomode, @@ -1855,19 +1895,17 @@ void pnfs_error_mark_layout_for_return(struct inode *inode, bool return_now = false; spin_lock(&inode->i_lock); - pnfs_set_plh_return_info(lo, range.iomode, lseg->pls_seq); + pnfs_set_plh_return_info(lo, range.iomode, 0); /* * mark all matching lsegs so that we are sure to have no live * segments at hand when sending layoutreturn. See pnfs_put_lseg() * for how it works. */ - if (!pnfs_mark_matching_lsegs_return(lo, &free_me, - &range, lseg->pls_seq)) { + if (!pnfs_mark_matching_lsegs_return(lo, &free_me, &range, 0)) { nfs4_stateid stateid; - enum pnfs_iomode iomode = lo->plh_return_iomode; + enum pnfs_iomode iomode; - nfs4_stateid_copy(&stateid, &lo->plh_stateid); - return_now = pnfs_prepare_layoutreturn(lo); + return_now = pnfs_prepare_layoutreturn(lo, &stateid, &iomode); spin_unlock(&inode->i_lock); if (return_now) pnfs_send_layoutreturn(lo, &stateid, iomode, false); @@ -2382,7 +2420,10 @@ pnfs_layoutcommit_inode(struct inode *inode, bool sync) nfs_fattr_init(&data->fattr); data->args.bitmask = NFS_SERVER(inode)->cache_consistency_bitmask; data->res.fattr = &data->fattr; - data->args.lastbytewritten = end_pos - 1; + if (end_pos != 0) + data->args.lastbytewritten = end_pos - 1; + else + data->args.lastbytewritten = U64_MAX; data->res.server = NFS_SERVER(inode); if (ld->prepare_layoutcommit) { diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h index b21bd0bee784..31d99b2927b0 100644 --- a/fs/nfs/pnfs.h +++ b/fs/nfs/pnfs.h @@ -268,6 +268,8 @@ int pnfs_mark_matching_lsegs_return(struct pnfs_layout_hdr *lo, struct list_head *tmp_list, const struct pnfs_layout_range *recall_range, u32 seq); +int pnfs_mark_layout_stateid_invalid(struct pnfs_layout_hdr *lo, + struct list_head *lseg_list); bool pnfs_roc(struct inode *ino); void pnfs_roc_release(struct inode *ino); void pnfs_roc_set_barrier(struct inode *ino, u32 barrier); @@ -375,6 +377,11 @@ static inline bool nfs_have_layout(struct inode *inode) return NFS_I(inode)->layout != NULL; } +static inline bool pnfs_layout_is_valid(const struct pnfs_layout_hdr *lo) +{ + return test_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags) == 0; +} + static inline struct nfs4_deviceid_node * nfs4_get_deviceid(struct nfs4_deviceid_node *d) { @@ -545,19 +552,6 @@ pnfs_calc_offset_length(u64 offset, u64 end) return 1 + end - offset; } -/** - * pnfs_mark_layout_returned_if_empty - marks the layout as returned - * @lo: layout header - * - * Note: Caller must hold inode->i_lock - */ -static inline void -pnfs_mark_layout_returned_if_empty(struct pnfs_layout_hdr *lo) -{ - if (list_empty(&lo->plh_segs)) - set_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags); -} - static inline void pnfs_copy_range(struct pnfs_layout_range *dst, const struct pnfs_layout_range *src) @@ -629,6 +623,13 @@ pnfs_sync_inode(struct inode *inode, bool datasync) } static inline bool +pnfs_layoutcommit_outstanding(struct inode *inode) +{ + return false; +} + + +static inline bool pnfs_roc(struct inode *ino) { return false; @@ -716,13 +717,6 @@ pnfs_use_threshold(struct nfs4_threshold **dst, struct nfs4_threshold *src, return false; } -static inline bool -pnfs_layoutcommit_outstanding(struct inode *inode) -{ - return false; -} - - static inline struct nfs4_threshold *pnfs_mdsthreshold_alloc(void) { return NULL; diff --git a/fs/nfs/pnfs_nfs.c b/fs/nfs/pnfs_nfs.c index b38e3c0dc790..f3468b57a32a 100644 --- a/fs/nfs/pnfs_nfs.c +++ b/fs/nfs/pnfs_nfs.c @@ -595,7 +595,7 @@ static void nfs4_clear_ds_conn_bit(struct nfs4_pnfs_ds *ds) } static struct nfs_client *(*get_v3_ds_connect)( - struct nfs_client *mds_clp, + struct nfs_server *mds_srv, const struct sockaddr *ds_addr, int ds_addrlen, int ds_proto, @@ -654,7 +654,7 @@ static int _nfs4_pnfs_v3_ds_connect(struct nfs_server *mds_srv, rpc_clnt_add_xprt(clp->cl_rpcclient, &xprt_args, rpc_clnt_test_and_add_xprt, NULL); } else - clp = get_v3_ds_connect(mds_srv->nfs_client, + clp = get_v3_ds_connect(mds_srv, (struct sockaddr *)&da->da_addr, da->da_addrlen, IPPROTO_TCP, timeo, retrans, au_flavor); @@ -690,7 +690,7 @@ static int _nfs4_pnfs_v4_ds_connect(struct nfs_server *mds_srv, dprintk("%s: DS %s: trying address %s\n", __func__, ds->ds_remotestr, da->da_remotestr); - clp = nfs4_set_ds_client(mds_srv->nfs_client, + clp = nfs4_set_ds_client(mds_srv, (struct sockaddr *)&da->da_addr, da->da_addrlen, IPPROTO_TCP, timeo, retrans, minor_version, @@ -940,6 +940,13 @@ EXPORT_SYMBOL_GPL(pnfs_layout_mark_request_commit); int pnfs_nfs_generic_sync(struct inode *inode, bool datasync) { + int ret; + + if (!pnfs_layoutcommit_outstanding(inode)) + return 0; + ret = nfs_commit_inode(inode, FLUSH_SYNC); + if (ret < 0) + return ret; if (datasync) return 0; return pnfs_layoutcommit_inode(inode, true); diff --git a/fs/nfs/super.c b/fs/nfs/super.c index 2137e0202f25..18d446e1a82b 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -1684,6 +1684,7 @@ static int nfs_verify_authflavors(struct nfs_parsed_mount_data *args, { rpc_authflavor_t flavor = RPC_AUTH_MAXFLAVOR; unsigned int i; + int use_auth_null = false; /* * If the sec= mount option is used, the specified flavor or AUTH_NULL @@ -1691,14 +1692,21 @@ static int nfs_verify_authflavors(struct nfs_parsed_mount_data *args, * * AUTH_NULL has a special meaning when it's in the server list - it * means that the server will ignore the rpc creds, so any flavor - * can be used. + * can be used but still use the sec= that was specified. */ for (i = 0; i < count; i++) { flavor = server_authlist[i]; - if (nfs_auth_info_match(&args->auth_info, flavor) || - flavor == RPC_AUTH_NULL) + if (nfs_auth_info_match(&args->auth_info, flavor)) goto out; + + if (flavor == RPC_AUTH_NULL) + use_auth_null = true; + } + + if (use_auth_null) { + flavor = RPC_AUTH_NULL; + goto out; } dfprintk(MOUNT, diff --git a/fs/nfs/write.c b/fs/nfs/write.c index 593fa21a02c0..3a6724c6eb5f 100644 --- a/fs/nfs/write.c +++ b/fs/nfs/write.c @@ -625,7 +625,7 @@ static int nfs_writepage_locked(struct page *page, int err; nfs_inc_stats(inode, NFSIOS_VFSWRITEPAGE); - nfs_pageio_init_write(&pgio, inode, wb_priority(wbc), + nfs_pageio_init_write(&pgio, inode, 0, false, &nfs_async_write_completion_ops); err = nfs_do_writepage(page, wbc, &pgio, launder); nfs_pageio_complete(&pgio); @@ -657,16 +657,9 @@ static int nfs_writepages_callback(struct page *page, struct writeback_control * int nfs_writepages(struct address_space *mapping, struct writeback_control *wbc) { struct inode *inode = mapping->host; - unsigned long *bitlock = &NFS_I(inode)->flags; struct nfs_pageio_descriptor pgio; int err; - /* Stop dirtying of new pages while we sync */ - err = wait_on_bit_lock_action(bitlock, NFS_INO_FLUSHING, - nfs_wait_bit_killable, TASK_KILLABLE); - if (err) - goto out_err; - nfs_inc_stats(inode, NFSIOS_VFSWRITEPAGES); nfs_pageio_init_write(&pgio, inode, wb_priority(wbc), false, @@ -674,10 +667,6 @@ int nfs_writepages(struct address_space *mapping, struct writeback_control *wbc) err = write_cache_pages(mapping, wbc, nfs_writepages_callback, &pgio); nfs_pageio_complete(&pgio); - clear_bit_unlock(NFS_INO_FLUSHING, bitlock); - smp_mb__after_atomic(); - wake_up_bit(bitlock, NFS_INO_FLUSHING); - if (err < 0) goto out_err; err = pgio.pg_error; @@ -1195,9 +1184,11 @@ nfs_key_timeout_notify(struct file *filp, struct inode *inode) /* * Test if the open context credential key is marked to expire soon. */ -bool nfs_ctx_key_to_expire(struct nfs_open_context *ctx) +bool nfs_ctx_key_to_expire(struct nfs_open_context *ctx, struct inode *inode) { - return rpcauth_cred_key_to_expire(ctx->cred); + struct rpc_auth *auth = NFS_SERVER(inode)->client->cl_auth; + + return rpcauth_cred_key_to_expire(auth, ctx->cred); } /* @@ -1289,6 +1280,9 @@ int nfs_updatepage(struct file *file, struct page *page, dprintk("NFS: nfs_updatepage(%pD2 %d@%lld)\n", file, count, (long long)(page_file_offset(page) + offset)); + if (!count) + goto out; + if (nfs_can_extend_write(file, page, inode)) { count = max(count + offset, nfs_page_length(page)); offset = 0; @@ -1299,7 +1293,7 @@ int nfs_updatepage(struct file *file, struct page *page, nfs_set_pageerror(page); else __set_page_dirty_nobuffers(page); - +out: dprintk("NFS: nfs_updatepage returns %d (isize %lld)\n", status, (long long)i_size_read(inode)); return status; @@ -1800,7 +1794,7 @@ static void nfs_commit_release_pages(struct nfs_commit_data *data) /* Okay, COMMIT succeeded, apparently. Check the verifier * returned by the server against all stored verfs. */ - if (!memcmp(&req->wb_verf, &data->verf.verifier, sizeof(req->wb_verf))) { + if (!nfs_write_verifier_cmp(&req->wb_verf, &data->verf.verifier)) { /* We have a match */ nfs_inode_remove_request(req); dprintk(" OK\n"); @@ -1924,6 +1918,24 @@ out_mark_dirty: EXPORT_SYMBOL_GPL(nfs_write_inode); /* + * Wrapper for filemap_write_and_wait_range() + * + * Needed for pNFS in order to ensure data becomes visible to the + * client. + */ +int nfs_filemap_write_and_wait_range(struct address_space *mapping, + loff_t lstart, loff_t lend) +{ + int ret; + + ret = filemap_write_and_wait_range(mapping, lstart, lend); + if (ret == 0) + ret = pnfs_sync_inode(mapping->host, true); + return ret; +} +EXPORT_SYMBOL_GPL(nfs_filemap_write_and_wait_range); + +/* * flush the inode to disk. */ int nfs_wb_all(struct inode *inode) diff --git a/fs/nilfs2/alloc.c b/fs/nilfs2/alloc.c index 1a85d94f5b25..2c90e285d7c6 100644 --- a/fs/nilfs2/alloc.c +++ b/fs/nilfs2/alloc.c @@ -622,10 +622,10 @@ void nilfs_palloc_commit_free_entry(struct inode *inode, lock = nilfs_mdt_bgl_lock(inode, group); if (!nilfs_clear_bit_atomic(lock, group_offset, bitmap)) - nilfs_warning(inode->i_sb, __func__, - "entry number %llu already freed: ino=%lu", - (unsigned long long)req->pr_entry_nr, - (unsigned long)inode->i_ino); + nilfs_msg(inode->i_sb, KERN_WARNING, + "%s (ino=%lu): entry number %llu already freed", + __func__, inode->i_ino, + (unsigned long long)req->pr_entry_nr); else nilfs_palloc_group_desc_add_entries(desc, lock, 1); @@ -663,10 +663,10 @@ void nilfs_palloc_abort_alloc_entry(struct inode *inode, lock = nilfs_mdt_bgl_lock(inode, group); if (!nilfs_clear_bit_atomic(lock, group_offset, bitmap)) - nilfs_warning(inode->i_sb, __func__, - "entry number %llu already freed: ino=%lu", - (unsigned long long)req->pr_entry_nr, - (unsigned long)inode->i_ino); + nilfs_msg(inode->i_sb, KERN_WARNING, + "%s (ino=%lu): entry number %llu already freed", + __func__, inode->i_ino, + (unsigned long long)req->pr_entry_nr); else nilfs_palloc_group_desc_add_entries(desc, lock, 1); @@ -772,10 +772,10 @@ int nilfs_palloc_freev(struct inode *inode, __u64 *entry_nrs, size_t nitems) do { if (!nilfs_clear_bit_atomic(lock, group_offset, bitmap)) { - nilfs_warning(inode->i_sb, __func__, - "entry number %llu already freed: ino=%lu", - (unsigned long long)entry_nrs[j], - (unsigned long)inode->i_ino); + nilfs_msg(inode->i_sb, KERN_WARNING, + "%s (ino=%lu): entry number %llu already freed", + __func__, inode->i_ino, + (unsigned long long)entry_nrs[j]); } else { n++; } @@ -816,12 +816,11 @@ int nilfs_palloc_freev(struct inode *inode, __u64 *entry_nrs, size_t nitems) for (k = 0; k < nempties; k++) { ret = nilfs_palloc_delete_entry_block(inode, last_nrs[k]); - if (ret && ret != -ENOENT) { - nilfs_warning(inode->i_sb, __func__, - "failed to delete block of entry %llu: ino=%lu, err=%d", - (unsigned long long)last_nrs[k], - (unsigned long)inode->i_ino, ret); - } + if (ret && ret != -ENOENT) + nilfs_msg(inode->i_sb, KERN_WARNING, + "error %d deleting block that object (entry=%llu, ino=%lu) belongs to", + ret, (unsigned long long)last_nrs[k], + inode->i_ino); } desc_kaddr = kmap_atomic(desc_bh->b_page); @@ -835,12 +834,10 @@ int nilfs_palloc_freev(struct inode *inode, __u64 *entry_nrs, size_t nitems) if (nfree == nilfs_palloc_entries_per_group(inode)) { ret = nilfs_palloc_delete_bitmap_block(inode, group); - if (ret && ret != -ENOENT) { - nilfs_warning(inode->i_sb, __func__, - "failed to delete bitmap block of group %lu: ino=%lu, err=%d", - group, - (unsigned long)inode->i_ino, ret); - } + if (ret && ret != -ENOENT) + nilfs_msg(inode->i_sb, KERN_WARNING, + "error %d deleting bitmap block of group=%lu, ino=%lu", + ret, group, inode->i_ino); } } return 0; diff --git a/fs/nilfs2/bmap.c b/fs/nilfs2/bmap.c index f2a7877e0c8c..01fb1831ca25 100644 --- a/fs/nilfs2/bmap.c +++ b/fs/nilfs2/bmap.c @@ -41,8 +41,8 @@ static int nilfs_bmap_convert_error(struct nilfs_bmap *bmap, struct inode *inode = bmap->b_inode; if (err == -EINVAL) { - nilfs_error(inode->i_sb, fname, - "broken bmap (inode number=%lu)", inode->i_ino); + __nilfs_error(inode->i_sb, fname, + "broken bmap (inode number=%lu)", inode->i_ino); err = -EIO; } return err; diff --git a/fs/nilfs2/bmap.h b/fs/nilfs2/bmap.h index b6a4c8f93ac8..2b6ffbe5997a 100644 --- a/fs/nilfs2/bmap.h +++ b/fs/nilfs2/bmap.h @@ -22,7 +22,7 @@ #include <linux/types.h> #include <linux/fs.h> #include <linux/buffer_head.h> -#include <linux/nilfs2_fs.h> +#include <linux/nilfs2_ondisk.h> /* nilfs_binfo, nilfs_inode, etc */ #include "alloc.h" #include "dat.h" diff --git a/fs/nilfs2/btnode.c b/fs/nilfs2/btnode.c index 4cca998ec7a0..d5c23da43513 100644 --- a/fs/nilfs2/btnode.c +++ b/fs/nilfs2/btnode.c @@ -41,7 +41,7 @@ nilfs_btnode_create_block(struct address_space *btnc, __u64 blocknr) struct inode *inode = NILFS_BTNC_I(btnc); struct buffer_head *bh; - bh = nilfs_grab_buffer(inode, btnc, blocknr, 1 << BH_NILFS_Node); + bh = nilfs_grab_buffer(inode, btnc, blocknr, BIT(BH_NILFS_Node)); if (unlikely(!bh)) return NULL; @@ -70,7 +70,7 @@ int nilfs_btnode_submit_block(struct address_space *btnc, __u64 blocknr, struct page *page; int err; - bh = nilfs_grab_buffer(inode, btnc, blocknr, 1 << BH_NILFS_Node); + bh = nilfs_grab_buffer(inode, btnc, blocknr, BIT(BH_NILFS_Node)); if (unlikely(!bh)) return -ENOMEM; diff --git a/fs/nilfs2/btree.c b/fs/nilfs2/btree.c index 982d1e3df3a5..2e315f9f2e51 100644 --- a/fs/nilfs2/btree.c +++ b/fs/nilfs2/btree.c @@ -339,12 +339,14 @@ static int nilfs_btree_node_lookup(const struct nilfs_btree_node *node, * nilfs_btree_node_broken - verify consistency of btree node * @node: btree node block to be examined * @size: node size (in bytes) + * @inode: host inode of btree * @blocknr: block number * * Return Value: If node is broken, 1 is returned. Otherwise, 0 is returned. */ static int nilfs_btree_node_broken(const struct nilfs_btree_node *node, - size_t size, sector_t blocknr) + size_t size, struct inode *inode, + sector_t blocknr) { int level, flags, nchildren; int ret = 0; @@ -358,9 +360,10 @@ static int nilfs_btree_node_broken(const struct nilfs_btree_node *node, (flags & NILFS_BTREE_NODE_ROOT) || nchildren < 0 || nchildren > NILFS_BTREE_NODE_NCHILDREN_MAX(size))) { - printk(KERN_CRIT "NILFS: bad btree node (blocknr=%llu): " - "level = %d, flags = 0x%x, nchildren = %d\n", - (unsigned long long)blocknr, level, flags, nchildren); + nilfs_msg(inode->i_sb, KERN_CRIT, + "bad btree node (ino=%lu, blocknr=%llu): level = %d, flags = 0x%x, nchildren = %d", + inode->i_ino, (unsigned long long)blocknr, level, + flags, nchildren); ret = 1; } return ret; @@ -369,12 +372,12 @@ static int nilfs_btree_node_broken(const struct nilfs_btree_node *node, /** * nilfs_btree_root_broken - verify consistency of btree root node * @node: btree root node to be examined - * @ino: inode number + * @inode: host inode of btree * * Return Value: If node is broken, 1 is returned. Otherwise, 0 is returned. */ static int nilfs_btree_root_broken(const struct nilfs_btree_node *node, - unsigned long ino) + struct inode *inode) { int level, flags, nchildren; int ret = 0; @@ -387,8 +390,9 @@ static int nilfs_btree_root_broken(const struct nilfs_btree_node *node, level >= NILFS_BTREE_LEVEL_MAX || nchildren < 0 || nchildren > NILFS_BTREE_ROOT_NCHILDREN_MAX)) { - pr_crit("NILFS: bad btree root (inode number=%lu): level = %d, flags = 0x%x, nchildren = %d\n", - ino, level, flags, nchildren); + nilfs_msg(inode->i_sb, KERN_CRIT, + "bad btree root (ino=%lu): level = %d, flags = 0x%x, nchildren = %d", + inode->i_ino, level, flags, nchildren); ret = 1; } return ret; @@ -396,13 +400,15 @@ static int nilfs_btree_root_broken(const struct nilfs_btree_node *node, int nilfs_btree_broken_node_block(struct buffer_head *bh) { + struct inode *inode; int ret; if (buffer_nilfs_checked(bh)) return 0; + inode = bh->b_page->mapping->host; ret = nilfs_btree_node_broken((struct nilfs_btree_node *)bh->b_data, - bh->b_size, bh->b_blocknr); + bh->b_size, inode, bh->b_blocknr); if (likely(!ret)) set_buffer_nilfs_checked(bh); return ret; @@ -448,13 +454,15 @@ nilfs_btree_get_node(const struct nilfs_bmap *btree, return node; } -static int -nilfs_btree_bad_node(struct nilfs_btree_node *node, int level) +static int nilfs_btree_bad_node(const struct nilfs_bmap *btree, + struct nilfs_btree_node *node, int level) { if (unlikely(nilfs_btree_node_get_level(node) != level)) { dump_stack(); - printk(KERN_CRIT "NILFS: btree level mismatch: %d != %d\n", - nilfs_btree_node_get_level(node), level); + nilfs_msg(btree->b_inode->i_sb, KERN_CRIT, + "btree level mismatch (ino=%lu): %d != %d", + btree->b_inode->i_ino, + nilfs_btree_node_get_level(node), level); return 1; } return 0; @@ -509,6 +517,9 @@ static int __nilfs_btree_get_block(const struct nilfs_bmap *btree, __u64 ptr, out_no_wait: if (!buffer_uptodate(bh)) { + nilfs_msg(btree->b_inode->i_sb, KERN_ERR, + "I/O error reading b-tree node block (ino=%lu, blocknr=%llu)", + btree->b_inode->i_ino, (unsigned long long)ptr); brelse(bh); return -EIO; } @@ -568,7 +579,7 @@ static int nilfs_btree_do_lookup(const struct nilfs_bmap *btree, return ret; node = nilfs_btree_get_nonroot_node(path, level); - if (nilfs_btree_bad_node(node, level)) + if (nilfs_btree_bad_node(btree, node, level)) return -EINVAL; if (!found) found = nilfs_btree_node_lookup(node, key, &index); @@ -616,7 +627,7 @@ static int nilfs_btree_do_lookup_last(const struct nilfs_bmap *btree, if (ret < 0) return ret; node = nilfs_btree_get_nonroot_node(path, level); - if (nilfs_btree_bad_node(node, level)) + if (nilfs_btree_bad_node(btree, node, level)) return -EINVAL; index = nilfs_btree_node_get_nchildren(node) - 1; ptr = nilfs_btree_node_get_ptr(node, index, ncmax); @@ -2072,8 +2083,10 @@ static int nilfs_btree_propagate(struct nilfs_bmap *btree, ret = nilfs_btree_do_lookup(btree, path, key, NULL, level + 1, 0); if (ret < 0) { if (unlikely(ret == -ENOENT)) - printk(KERN_CRIT "%s: key = %llu, level == %d\n", - __func__, (unsigned long long)key, level); + nilfs_msg(btree->b_inode->i_sb, KERN_CRIT, + "writing node/leaf block does not appear in b-tree (ino=%lu) at key=%llu, level=%d", + btree->b_inode->i_ino, + (unsigned long long)key, level); goto out; } @@ -2110,12 +2123,11 @@ static void nilfs_btree_add_dirty_buffer(struct nilfs_bmap *btree, if (level < NILFS_BTREE_LEVEL_NODE_MIN || level >= NILFS_BTREE_LEVEL_MAX) { dump_stack(); - printk(KERN_WARNING - "%s: invalid btree level: %d (key=%llu, ino=%lu, " - "blocknr=%llu)\n", - __func__, level, (unsigned long long)key, - NILFS_BMAP_I(btree)->vfs_inode.i_ino, - (unsigned long long)bh->b_blocknr); + nilfs_msg(btree->b_inode->i_sb, KERN_WARNING, + "invalid btree level: %d (key=%llu, ino=%lu, blocknr=%llu)", + level, (unsigned long long)key, + btree->b_inode->i_ino, + (unsigned long long)bh->b_blocknr); return; } @@ -2394,8 +2406,7 @@ int nilfs_btree_init(struct nilfs_bmap *bmap) __nilfs_btree_init(bmap); - if (nilfs_btree_root_broken(nilfs_btree_get_root(bmap), - bmap->b_inode->i_ino)) + if (nilfs_btree_root_broken(nilfs_btree_get_root(bmap), bmap->b_inode)) ret = -EIO; return ret; } diff --git a/fs/nilfs2/btree.h b/fs/nilfs2/btree.h index df1a25faa83b..2184e47fa4bf 100644 --- a/fs/nilfs2/btree.h +++ b/fs/nilfs2/btree.h @@ -22,7 +22,7 @@ #include <linux/types.h> #include <linux/buffer_head.h> #include <linux/list.h> -#include <linux/nilfs2_fs.h> +#include <linux/nilfs2_ondisk.h> /* nilfs_btree_node */ #include "btnode.h" #include "bmap.h" diff --git a/fs/nilfs2/cpfile.c b/fs/nilfs2/cpfile.c index 8a3d3b65af3f..a15a1601e931 100644 --- a/fs/nilfs2/cpfile.c +++ b/fs/nilfs2/cpfile.c @@ -21,7 +21,6 @@ #include <linux/string.h> #include <linux/buffer_head.h> #include <linux/errno.h> -#include <linux/nilfs2_fs.h> #include "mdt.h" #include "cpfile.h" @@ -332,9 +331,9 @@ int nilfs_cpfile_delete_checkpoints(struct inode *cpfile, int ret, ncps, nicps, nss, count, i; if (unlikely(start == 0 || start > end)) { - printk(KERN_ERR "%s: invalid range of checkpoint numbers: " - "[%llu, %llu)\n", __func__, - (unsigned long long)start, (unsigned long long)end); + nilfs_msg(cpfile->i_sb, KERN_ERR, + "cannot delete checkpoints: invalid range [%llu, %llu)", + (unsigned long long)start, (unsigned long long)end); return -EINVAL; } @@ -386,9 +385,9 @@ int nilfs_cpfile_delete_checkpoints(struct inode *cpfile, cpfile, cno); if (ret == 0) continue; - printk(KERN_ERR - "%s: cannot delete block\n", - __func__); + nilfs_msg(cpfile->i_sb, KERN_ERR, + "error %d deleting checkpoint block", + ret); break; } } @@ -991,14 +990,12 @@ int nilfs_cpfile_read(struct super_block *sb, size_t cpsize, int err; if (cpsize > sb->s_blocksize) { - printk(KERN_ERR - "NILFS: too large checkpoint size: %zu bytes.\n", - cpsize); + nilfs_msg(sb, KERN_ERR, + "too large checkpoint size: %zu bytes", cpsize); return -EINVAL; } else if (cpsize < NILFS_MIN_CHECKPOINT_SIZE) { - printk(KERN_ERR - "NILFS: too small checkpoint size: %zu bytes.\n", - cpsize); + nilfs_msg(sb, KERN_ERR, + "too small checkpoint size: %zu bytes", cpsize); return -EINVAL; } diff --git a/fs/nilfs2/cpfile.h b/fs/nilfs2/cpfile.h index 0249744ae234..6eca972f9673 100644 --- a/fs/nilfs2/cpfile.h +++ b/fs/nilfs2/cpfile.h @@ -21,7 +21,8 @@ #include <linux/fs.h> #include <linux/buffer_head.h> -#include <linux/nilfs2_fs.h> +#include <linux/nilfs2_api.h> /* nilfs_cpstat */ +#include <linux/nilfs2_ondisk.h> /* nilfs_inode, nilfs_checkpoint */ int nilfs_cpfile_get_checkpoint(struct inode *, __u64, int, diff --git a/fs/nilfs2/dat.c b/fs/nilfs2/dat.c index 7367610ea807..dffedb2f8817 100644 --- a/fs/nilfs2/dat.c +++ b/fs/nilfs2/dat.c @@ -349,10 +349,11 @@ int nilfs_dat_move(struct inode *dat, __u64 vblocknr, sector_t blocknr) kaddr = kmap_atomic(entry_bh->b_page); entry = nilfs_palloc_block_get_entry(dat, vblocknr, entry_bh, kaddr); if (unlikely(entry->de_blocknr == cpu_to_le64(0))) { - printk(KERN_CRIT "%s: vbn = %llu, [%llu, %llu)\n", __func__, - (unsigned long long)vblocknr, - (unsigned long long)le64_to_cpu(entry->de_start), - (unsigned long long)le64_to_cpu(entry->de_end)); + nilfs_msg(dat->i_sb, KERN_CRIT, + "%s: invalid vblocknr = %llu, [%llu, %llu)", + __func__, (unsigned long long)vblocknr, + (unsigned long long)le64_to_cpu(entry->de_start), + (unsigned long long)le64_to_cpu(entry->de_end)); kunmap_atomic(kaddr); brelse(entry_bh); return -EINVAL; @@ -479,14 +480,12 @@ int nilfs_dat_read(struct super_block *sb, size_t entry_size, int err; if (entry_size > sb->s_blocksize) { - printk(KERN_ERR - "NILFS: too large DAT entry size: %zu bytes.\n", - entry_size); + nilfs_msg(sb, KERN_ERR, "too large DAT entry size: %zu bytes", + entry_size); return -EINVAL; } else if (entry_size < NILFS_MIN_DAT_ENTRY_SIZE) { - printk(KERN_ERR - "NILFS: too small DAT entry size: %zu bytes.\n", - entry_size); + nilfs_msg(sb, KERN_ERR, "too small DAT entry size: %zu bytes", + entry_size); return -EINVAL; } diff --git a/fs/nilfs2/dat.h b/fs/nilfs2/dat.h index abbfdabcabea..57dc6cf466d0 100644 --- a/fs/nilfs2/dat.h +++ b/fs/nilfs2/dat.h @@ -22,6 +22,7 @@ #include <linux/types.h> #include <linux/buffer_head.h> #include <linux/fs.h> +#include <linux/nilfs2_ondisk.h> /* nilfs_inode, nilfs_checkpoint */ struct nilfs_palloc_req; diff --git a/fs/nilfs2/dir.c b/fs/nilfs2/dir.c index e506f4f7120a..908ebbf0ac7e 100644 --- a/fs/nilfs2/dir.c +++ b/fs/nilfs2/dir.c @@ -42,6 +42,28 @@ #include "nilfs.h" #include "page.h" +static inline unsigned int nilfs_rec_len_from_disk(__le16 dlen) +{ + unsigned int len = le16_to_cpu(dlen); + +#if (PAGE_SIZE >= 65536) + if (len == NILFS_MAX_REC_LEN) + return 1 << 16; +#endif + return len; +} + +static inline __le16 nilfs_rec_len_to_disk(unsigned int len) +{ +#if (PAGE_SIZE >= 65536) + if (len == (1 << 16)) + return cpu_to_le16(NILFS_MAX_REC_LEN); + + BUG_ON(len > (1 << 16)); +#endif + return cpu_to_le16(len); +} + /* * nilfs uses block-sized chunks. Arguably, sector-sized ones would be * more robust, but we have what we have @@ -140,10 +162,9 @@ out: /* Too bad, we had an error */ Ebadsize: - nilfs_error(sb, "nilfs_check_page", + nilfs_error(sb, "size of directory #%lu is not a multiple of chunk size", - dir->i_ino - ); + dir->i_ino); goto fail; Eshort: error = "rec_len is smaller than minimal"; @@ -157,19 +178,18 @@ Enamelen: Espan: error = "directory entry across blocks"; bad_entry: - nilfs_error(sb, "nilfs_check_page", "bad entry in directory #%lu: %s - " - "offset=%lu, inode=%lu, rec_len=%d, name_len=%d", - dir->i_ino, error, (page->index<<PAGE_SHIFT)+offs, - (unsigned long) le64_to_cpu(p->inode), + nilfs_error(sb, + "bad entry in directory #%lu: %s - offset=%lu, inode=%lu, rec_len=%d, name_len=%d", + dir->i_ino, error, (page->index << PAGE_SHIFT) + offs, + (unsigned long)le64_to_cpu(p->inode), rec_len, p->name_len); goto fail; Eend: p = (struct nilfs_dir_entry *)(kaddr + offs); - nilfs_error(sb, "nilfs_check_page", - "entry in directory #%lu spans the page boundary" - "offset=%lu, inode=%lu", - dir->i_ino, (page->index<<PAGE_SHIFT)+offs, - (unsigned long) le64_to_cpu(p->inode)); + nilfs_error(sb, + "entry in directory #%lu spans the page boundary offset=%lu, inode=%lu", + dir->i_ino, (page->index << PAGE_SHIFT) + offs, + (unsigned long)le64_to_cpu(p->inode)); fail: SetPageError(page); return false; @@ -267,8 +287,7 @@ static int nilfs_readdir(struct file *file, struct dir_context *ctx) struct page *page = nilfs_get_page(inode, n); if (IS_ERR(page)) { - nilfs_error(sb, __func__, "bad page in #%lu", - inode->i_ino); + nilfs_error(sb, "bad page in #%lu", inode->i_ino); ctx->pos += PAGE_SIZE - offset; return -EIO; } @@ -278,8 +297,7 @@ static int nilfs_readdir(struct file *file, struct dir_context *ctx) NILFS_DIR_REC_LEN(1); for ( ; (char *)de <= limit; de = nilfs_next_entry(de)) { if (de->rec_len == 0) { - nilfs_error(sb, __func__, - "zero-length directory entry"); + nilfs_error(sb, "zero-length directory entry"); nilfs_put_page(page); return -EIO; } @@ -345,7 +363,7 @@ nilfs_find_entry(struct inode *dir, const struct qstr *qstr, kaddr += nilfs_last_byte(dir, n) - reclen; while ((char *) de <= kaddr) { if (de->rec_len == 0) { - nilfs_error(dir->i_sb, __func__, + nilfs_error(dir->i_sb, "zero-length directory entry"); nilfs_put_page(page); goto out; @@ -360,7 +378,7 @@ nilfs_find_entry(struct inode *dir, const struct qstr *qstr, n = 0; /* next page is past the blocks we've got */ if (unlikely(n > (dir->i_blocks >> (PAGE_SHIFT - 9)))) { - nilfs_error(dir->i_sb, __func__, + nilfs_error(dir->i_sb, "dir %lu size %lld exceeds block count %llu", dir->i_ino, dir->i_size, (unsigned long long)dir->i_blocks); @@ -469,7 +487,7 @@ int nilfs_add_link(struct dentry *dentry, struct inode *inode) goto got_it; } if (de->rec_len == 0) { - nilfs_error(dir->i_sb, __func__, + nilfs_error(dir->i_sb, "zero-length directory entry"); err = -EIO; goto out_unlock; @@ -541,7 +559,7 @@ int nilfs_delete_entry(struct nilfs_dir_entry *dir, struct page *page) while ((char *)de < (char *)dir) { if (de->rec_len == 0) { - nilfs_error(inode->i_sb, __func__, + nilfs_error(inode->i_sb, "zero-length directory entry"); err = -EIO; goto out; @@ -628,7 +646,7 @@ int nilfs_empty_dir(struct inode *inode) while ((char *)de <= kaddr) { if (de->rec_len == 0) { - nilfs_error(inode->i_sb, __func__, + nilfs_error(inode->i_sb, "zero-length directory entry (kaddr=%p, de=%p)", kaddr, de); goto not_empty; diff --git a/fs/nilfs2/direct.c b/fs/nilfs2/direct.c index 251a44928405..96e3ed0d9652 100644 --- a/fs/nilfs2/direct.c +++ b/fs/nilfs2/direct.c @@ -337,14 +337,16 @@ static int nilfs_direct_assign(struct nilfs_bmap *bmap, key = nilfs_bmap_data_get_key(bmap, *bh); if (unlikely(key > NILFS_DIRECT_KEY_MAX)) { - printk(KERN_CRIT "%s: invalid key: %llu\n", __func__, - (unsigned long long)key); + nilfs_msg(bmap->b_inode->i_sb, KERN_CRIT, + "%s (ino=%lu): invalid key: %llu", __func__, + bmap->b_inode->i_ino, (unsigned long long)key); return -EINVAL; } ptr = nilfs_direct_get_ptr(bmap, key); if (unlikely(ptr == NILFS_BMAP_INVALID_PTR)) { - printk(KERN_CRIT "%s: invalid pointer: %llu\n", __func__, - (unsigned long long)ptr); + nilfs_msg(bmap->b_inode->i_sb, KERN_CRIT, + "%s (ino=%lu): invalid pointer: %llu", __func__, + bmap->b_inode->i_ino, (unsigned long long)ptr); return -EINVAL; } diff --git a/fs/nilfs2/direct.h b/fs/nilfs2/direct.h index 3015a6e78724..cfe85e848bba 100644 --- a/fs/nilfs2/direct.h +++ b/fs/nilfs2/direct.h @@ -24,16 +24,6 @@ #include "bmap.h" -/** - * struct nilfs_direct_node - direct node - * @dn_flags: flags - * @dn_pad: padding - */ -struct nilfs_direct_node { - __u8 dn_flags; - __u8 pad[7]; -}; - #define NILFS_DIRECT_NBLOCKS (NILFS_BMAP_SIZE / sizeof(__le64) - 1) #define NILFS_DIRECT_KEY_MIN 0 #define NILFS_DIRECT_KEY_MAX (NILFS_DIRECT_NBLOCKS - 1) diff --git a/fs/nilfs2/gcinode.c b/fs/nilfs2/gcinode.c index e9148f94d696..853a831dcde0 100644 --- a/fs/nilfs2/gcinode.c +++ b/fs/nilfs2/gcinode.c @@ -148,8 +148,15 @@ int nilfs_gccache_submit_read_node(struct inode *inode, sector_t pbn, int nilfs_gccache_wait_and_mark_dirty(struct buffer_head *bh) { wait_on_buffer(bh); - if (!buffer_uptodate(bh)) + if (!buffer_uptodate(bh)) { + struct inode *inode = bh->b_page->mapping->host; + + nilfs_msg(inode->i_sb, KERN_ERR, + "I/O error reading %s block for GC (ino=%lu, vblocknr=%llu)", + buffer_nilfs_node(bh) ? "node" : "data", + inode->i_ino, (unsigned long long)bh->b_blocknr); return -EIO; + } if (buffer_dirty(bh)) return -EEXIST; diff --git a/fs/nilfs2/ifile.c b/fs/nilfs2/ifile.c index 1d2b1805327a..b8fa45c20c63 100644 --- a/fs/nilfs2/ifile.c +++ b/fs/nilfs2/ifile.c @@ -145,15 +145,14 @@ int nilfs_ifile_get_inode_block(struct inode *ifile, ino_t ino, int err; if (unlikely(!NILFS_VALID_INODE(sb, ino))) { - nilfs_error(sb, __func__, "bad inode number: %lu", - (unsigned long) ino); + nilfs_error(sb, "bad inode number: %lu", (unsigned long)ino); return -EINVAL; } err = nilfs_palloc_get_entry_block(ifile, ino, 0, out_bh); if (unlikely(err)) - nilfs_warning(sb, __func__, "unable to read inode: %lu", - (unsigned long) ino); + nilfs_msg(sb, KERN_WARNING, "error %d reading inode: ino=%lu", + err, (unsigned long)ino); return err; } diff --git a/fs/nilfs2/ifile.h b/fs/nilfs2/ifile.h index 23ad2f091e76..188b94fe0ec5 100644 --- a/fs/nilfs2/ifile.h +++ b/fs/nilfs2/ifile.h @@ -23,7 +23,6 @@ #include <linux/fs.h> #include <linux/buffer_head.h> -#include <linux/nilfs2_fs.h> #include "mdt.h" #include "alloc.h" diff --git a/fs/nilfs2/inode.c b/fs/nilfs2/inode.c index a0ebdb17e912..af04f553d7c9 100644 --- a/fs/nilfs2/inode.c +++ b/fs/nilfs2/inode.c @@ -112,13 +112,10 @@ int nilfs_get_block(struct inode *inode, sector_t blkoff, * However, the page having this block must * be locked in this case. */ - printk(KERN_WARNING - "nilfs_get_block: a race condition " - "while inserting a data block. " - "(inode number=%lu, file block " - "offset=%llu)\n", - inode->i_ino, - (unsigned long long)blkoff); + nilfs_msg(inode->i_sb, KERN_WARNING, + "%s (ino=%lu): a race condition while inserting a data block at offset=%llu", + __func__, inode->i_ino, + (unsigned long long)blkoff); err = 0; } nilfs_transaction_abort(inode->i_sb); @@ -359,7 +356,7 @@ struct inode *nilfs_new_inode(struct inode *dir, umode_t mode) root = NILFS_I(dir)->i_root; ii = NILFS_I(inode); - ii->i_state = 1 << NILFS_I_NEW; + ii->i_state = BIT(NILFS_I_NEW); ii->i_root = root; err = nilfs_ifile_create_inode(root->ifile, &ino, &ii->i_bh); @@ -558,7 +555,7 @@ static int nilfs_iget_set(struct inode *inode, void *opaque) inode->i_ino = args->ino; if (args->for_gc) { - NILFS_I(inode)->i_state = 1 << NILFS_I_GCINODE; + NILFS_I(inode)->i_state = BIT(NILFS_I_GCINODE); NILFS_I(inode)->i_cno = args->cno; NILFS_I(inode)->i_root = NULL; } else { @@ -726,9 +723,9 @@ repeat: goto repeat; failed: - nilfs_warning(ii->vfs_inode.i_sb, __func__, - "failed to truncate bmap (ino=%lu, err=%d)", - ii->vfs_inode.i_ino, ret); + nilfs_msg(ii->vfs_inode.i_sb, KERN_WARNING, + "error %d truncating bmap (ino=%lu)", ret, + ii->vfs_inode.i_ino); } void nilfs_truncate(struct inode *inode) @@ -939,9 +936,9 @@ int nilfs_set_file_dirty(struct inode *inode, unsigned int nr_dirty) * This will happen when somebody is freeing * this inode. */ - nilfs_warning(inode->i_sb, __func__, - "cannot get inode (ino=%lu)", - inode->i_ino); + nilfs_msg(inode->i_sb, KERN_WARNING, + "cannot set file dirty (ino=%lu): the file is being freed", + inode->i_ino); spin_unlock(&nilfs->ns_inode_lock); return -EINVAL; /* * NILFS_I_DIRTY may remain for @@ -962,8 +959,9 @@ int __nilfs_mark_inode_dirty(struct inode *inode, int flags) err = nilfs_load_inode_block(inode, &ibh); if (unlikely(err)) { - nilfs_warning(inode->i_sb, __func__, - "failed to reget inode block."); + nilfs_msg(inode->i_sb, KERN_WARNING, + "cannot mark inode dirty (ino=%lu): error %d loading inode block", + inode->i_ino, err); return err; } nilfs_update_inode(inode, ibh, flags); @@ -989,8 +987,8 @@ void nilfs_dirty_inode(struct inode *inode, int flags) struct nilfs_mdt_info *mdi = NILFS_MDT(inode); if (is_bad_inode(inode)) { - nilfs_warning(inode->i_sb, __func__, - "tried to mark bad_inode dirty. ignored."); + nilfs_msg(inode->i_sb, KERN_WARNING, + "tried to mark bad_inode dirty. ignored."); dump_stack(); return; } diff --git a/fs/nilfs2/ioctl.c b/fs/nilfs2/ioctl.c index 358b57e2cdf9..f1d7989459fd 100644 --- a/fs/nilfs2/ioctl.c +++ b/fs/nilfs2/ioctl.c @@ -25,7 +25,6 @@ #include <linux/compat.h> /* compat_ptr() */ #include <linux/mount.h> /* mnt_want_write_file(), mnt_drop_write_file() */ #include <linux/buffer_head.h> -#include <linux/nilfs2_fs.h> #include "nilfs.h" #include "segment.h" #include "bmap.h" @@ -584,27 +583,25 @@ static int nilfs_ioctl_move_inode_block(struct inode *inode, if (unlikely(ret < 0)) { if (ret == -ENOENT) - printk(KERN_CRIT - "%s: invalid virtual block address (%s): " - "ino=%llu, cno=%llu, offset=%llu, " - "blocknr=%llu, vblocknr=%llu\n", - __func__, vdesc->vd_flags ? "node" : "data", - (unsigned long long)vdesc->vd_ino, - (unsigned long long)vdesc->vd_cno, - (unsigned long long)vdesc->vd_offset, - (unsigned long long)vdesc->vd_blocknr, - (unsigned long long)vdesc->vd_vblocknr); + nilfs_msg(inode->i_sb, KERN_CRIT, + "%s: invalid virtual block address (%s): ino=%llu, cno=%llu, offset=%llu, blocknr=%llu, vblocknr=%llu", + __func__, vdesc->vd_flags ? "node" : "data", + (unsigned long long)vdesc->vd_ino, + (unsigned long long)vdesc->vd_cno, + (unsigned long long)vdesc->vd_offset, + (unsigned long long)vdesc->vd_blocknr, + (unsigned long long)vdesc->vd_vblocknr); return ret; } if (unlikely(!list_empty(&bh->b_assoc_buffers))) { - printk(KERN_CRIT "%s: conflicting %s buffer: ino=%llu, " - "cno=%llu, offset=%llu, blocknr=%llu, vblocknr=%llu\n", - __func__, vdesc->vd_flags ? "node" : "data", - (unsigned long long)vdesc->vd_ino, - (unsigned long long)vdesc->vd_cno, - (unsigned long long)vdesc->vd_offset, - (unsigned long long)vdesc->vd_blocknr, - (unsigned long long)vdesc->vd_vblocknr); + nilfs_msg(inode->i_sb, KERN_CRIT, + "%s: conflicting %s buffer: ino=%llu, cno=%llu, offset=%llu, blocknr=%llu, vblocknr=%llu", + __func__, vdesc->vd_flags ? "node" : "data", + (unsigned long long)vdesc->vd_ino, + (unsigned long long)vdesc->vd_cno, + (unsigned long long)vdesc->vd_offset, + (unsigned long long)vdesc->vd_blocknr, + (unsigned long long)vdesc->vd_vblocknr); brelse(bh); return -EEXIST; } @@ -854,8 +851,8 @@ int nilfs_ioctl_prepare_clean_segments(struct the_nilfs *nilfs, return 0; failed: - printk(KERN_ERR "NILFS: GC failed during preparation: %s: err=%d\n", - msg, ret); + nilfs_msg(nilfs->ns_sb, KERN_ERR, "error %d preparing GC: %s", ret, + msg); return ret; } @@ -963,10 +960,11 @@ static int nilfs_ioctl_clean_segments(struct inode *inode, struct file *filp, } ret = nilfs_ioctl_move_blocks(inode->i_sb, &argv[0], kbufs[0]); - if (ret < 0) - printk(KERN_ERR "NILFS: GC failed during preparation: " - "cannot read source blocks: err=%d\n", ret); - else { + if (ret < 0) { + nilfs_msg(inode->i_sb, KERN_ERR, + "error %d preparing GC: cannot read source blocks", + ret); + } else { if (nilfs_sb_need_update(nilfs)) set_nilfs_discontinued(nilfs); ret = nilfs_clean_segments(inode->i_sb, argv, kbufs); diff --git a/fs/nilfs2/mdt.c b/fs/nilfs2/mdt.c index 0d7b71fbeff8..d56d3a5bea88 100644 --- a/fs/nilfs2/mdt.c +++ b/fs/nilfs2/mdt.c @@ -207,8 +207,12 @@ static int nilfs_mdt_read_block(struct inode *inode, unsigned long block, out_no_wait: err = -EIO; - if (!buffer_uptodate(first_bh)) + if (!buffer_uptodate(first_bh)) { + nilfs_msg(inode->i_sb, KERN_ERR, + "I/O error reading meta-data file (ino=%lu, block-offset=%lu)", + inode->i_ino, block); goto failed_bh; + } out: *out_bh = first_bh; return 0; diff --git a/fs/nilfs2/namei.c b/fs/nilfs2/namei.c index 1ec8ae5995a5..dbcf1dc93a51 100644 --- a/fs/nilfs2/namei.c +++ b/fs/nilfs2/namei.c @@ -283,9 +283,9 @@ static int nilfs_do_unlink(struct inode *dir, struct dentry *dentry) goto out; if (!inode->i_nlink) { - nilfs_warning(inode->i_sb, __func__, - "deleting nonexistent file (%lu), %d", - inode->i_ino, inode->i_nlink); + nilfs_msg(inode->i_sb, KERN_WARNING, + "deleting nonexistent file (ino=%lu), %d", + inode->i_ino, inode->i_nlink); set_nlink(inode, 1); } err = nilfs_delete_entry(de, page); diff --git a/fs/nilfs2/nilfs.h b/fs/nilfs2/nilfs.h index b1d48bc0532d..33f8c8fc96e8 100644 --- a/fs/nilfs2/nilfs.h +++ b/fs/nilfs2/nilfs.h @@ -23,7 +23,8 @@ #include <linux/buffer_head.h> #include <linux/spinlock.h> #include <linux/blkdev.h> -#include <linux/nilfs2_fs.h> +#include <linux/nilfs2_api.h> +#include <linux/nilfs2_ondisk.h> #include "the_nilfs.h" #include "bmap.h" @@ -119,20 +120,19 @@ enum { /* * Macros to check inode numbers */ -#define NILFS_MDT_INO_BITS \ - ((unsigned int)(1 << NILFS_DAT_INO | 1 << NILFS_CPFILE_INO | \ - 1 << NILFS_SUFILE_INO | 1 << NILFS_IFILE_INO | \ - 1 << NILFS_ATIME_INO | 1 << NILFS_SKETCH_INO)) +#define NILFS_MDT_INO_BITS \ + (BIT(NILFS_DAT_INO) | BIT(NILFS_CPFILE_INO) | \ + BIT(NILFS_SUFILE_INO) | BIT(NILFS_IFILE_INO) | \ + BIT(NILFS_ATIME_INO) | BIT(NILFS_SKETCH_INO)) -#define NILFS_SYS_INO_BITS \ - ((unsigned int)(1 << NILFS_ROOT_INO) | NILFS_MDT_INO_BITS) +#define NILFS_SYS_INO_BITS (BIT(NILFS_ROOT_INO) | NILFS_MDT_INO_BITS) #define NILFS_FIRST_INO(sb) (((struct the_nilfs *)sb->s_fs_info)->ns_first_ino) #define NILFS_MDT_INODE(sb, ino) \ - ((ino) < NILFS_FIRST_INO(sb) && (NILFS_MDT_INO_BITS & (1 << (ino)))) + ((ino) < NILFS_FIRST_INO(sb) && (NILFS_MDT_INO_BITS & BIT(ino))) #define NILFS_VALID_INODE(sb, ino) \ - ((ino) >= NILFS_FIRST_INO(sb) || (NILFS_SYS_INO_BITS & (1 << (ino)))) + ((ino) >= NILFS_FIRST_INO(sb) || (NILFS_SYS_INO_BITS & BIT(ino))) /** * struct nilfs_transaction_info: context information for synchronization @@ -299,10 +299,36 @@ static inline int nilfs_mark_inode_dirty_sync(struct inode *inode) /* super.c */ extern struct inode *nilfs_alloc_inode(struct super_block *); extern void nilfs_destroy_inode(struct inode *); + extern __printf(3, 4) -void nilfs_error(struct super_block *, const char *, const char *, ...); +void __nilfs_msg(struct super_block *sb, const char *level, + const char *fmt, ...); extern __printf(3, 4) -void nilfs_warning(struct super_block *, const char *, const char *, ...); +void __nilfs_error(struct super_block *sb, const char *function, + const char *fmt, ...); + +#ifdef CONFIG_PRINTK + +#define nilfs_msg(sb, level, fmt, ...) \ + __nilfs_msg(sb, level, fmt, ##__VA_ARGS__) +#define nilfs_error(sb, fmt, ...) \ + __nilfs_error(sb, __func__, fmt, ##__VA_ARGS__) + +#else + +#define nilfs_msg(sb, level, fmt, ...) \ + do { \ + no_printk(fmt, ##__VA_ARGS__); \ + (void)(sb); \ + } while (0) +#define nilfs_error(sb, fmt, ...) \ + do { \ + no_printk(fmt, ##__VA_ARGS__); \ + __nilfs_error(sb, "", " "); \ + } while (0) + +#endif /* CONFIG_PRINTK */ + extern struct nilfs_super_block * nilfs_read_super_block(struct super_block *, u64, int, struct buffer_head **); extern int nilfs_store_magic_and_option(struct super_block *, diff --git a/fs/nilfs2/page.c b/fs/nilfs2/page.c index d97ba5f11b77..f11a3ad2df0c 100644 --- a/fs/nilfs2/page.c +++ b/fs/nilfs2/page.c @@ -30,9 +30,9 @@ #include "mdt.h" -#define NILFS_BUFFER_INHERENT_BITS \ - ((1UL << BH_Uptodate) | (1UL << BH_Mapped) | (1UL << BH_NILFS_Node) | \ - (1UL << BH_NILFS_Volatile) | (1UL << BH_NILFS_Checked)) +#define NILFS_BUFFER_INHERENT_BITS \ + (BIT(BH_Uptodate) | BIT(BH_Mapped) | BIT(BH_NILFS_Node) | \ + BIT(BH_NILFS_Volatile) | BIT(BH_NILFS_Checked)) static struct buffer_head * __nilfs_get_page_block(struct page *page, unsigned long block, pgoff_t index, @@ -85,9 +85,9 @@ void nilfs_forget_buffer(struct buffer_head *bh) { struct page *page = bh->b_page; const unsigned long clear_bits = - (1 << BH_Uptodate | 1 << BH_Dirty | 1 << BH_Mapped | - 1 << BH_Async_Write | 1 << BH_NILFS_Volatile | - 1 << BH_NILFS_Checked | 1 << BH_NILFS_Redirected); + (BIT(BH_Uptodate) | BIT(BH_Dirty) | BIT(BH_Mapped) | + BIT(BH_Async_Write) | BIT(BH_NILFS_Volatile) | + BIT(BH_NILFS_Checked) | BIT(BH_NILFS_Redirected)); lock_buffer(bh); set_mask_bits(&bh->b_state, clear_bits, 0); @@ -124,17 +124,17 @@ void nilfs_copy_buffer(struct buffer_head *dbh, struct buffer_head *sbh) dbh->b_bdev = sbh->b_bdev; bh = dbh; - bits = sbh->b_state & ((1UL << BH_Uptodate) | (1UL << BH_Mapped)); + bits = sbh->b_state & (BIT(BH_Uptodate) | BIT(BH_Mapped)); while ((bh = bh->b_this_page) != dbh) { lock_buffer(bh); bits &= bh->b_state; unlock_buffer(bh); } - if (bits & (1UL << BH_Uptodate)) + if (bits & BIT(BH_Uptodate)) SetPageUptodate(dpage); else ClearPageUptodate(dpage); - if (bits & (1UL << BH_Mapped)) + if (bits & BIT(BH_Mapped)) SetPageMappedToDisk(dpage); else ClearPageMappedToDisk(dpage); @@ -215,7 +215,7 @@ static void nilfs_copy_page(struct page *dst, struct page *src, int copy_dirty) create_empty_buffers(dst, sbh->b_size, 0); if (copy_dirty) - mask |= (1UL << BH_Dirty); + mask |= BIT(BH_Dirty); dbh = dbufs = page_buffers(dst); do { @@ -403,11 +403,10 @@ void nilfs_clear_dirty_page(struct page *page, bool silent) BUG_ON(!PageLocked(page)); - if (!silent) { - nilfs_warning(sb, __func__, - "discard page: offset %lld, ino %lu", - page_offset(page), inode->i_ino); - } + if (!silent) + nilfs_msg(sb, KERN_WARNING, + "discard dirty page: offset=%lld, ino=%lu", + page_offset(page), inode->i_ino); ClearPageUptodate(page); ClearPageMappedToDisk(page); @@ -415,18 +414,18 @@ void nilfs_clear_dirty_page(struct page *page, bool silent) if (page_has_buffers(page)) { struct buffer_head *bh, *head; const unsigned long clear_bits = - (1 << BH_Uptodate | 1 << BH_Dirty | 1 << BH_Mapped | - 1 << BH_Async_Write | 1 << BH_NILFS_Volatile | - 1 << BH_NILFS_Checked | 1 << BH_NILFS_Redirected); + (BIT(BH_Uptodate) | BIT(BH_Dirty) | BIT(BH_Mapped) | + BIT(BH_Async_Write) | BIT(BH_NILFS_Volatile) | + BIT(BH_NILFS_Checked) | BIT(BH_NILFS_Redirected)); bh = head = page_buffers(page); do { lock_buffer(bh); - if (!silent) { - nilfs_warning(sb, __func__, - "discard block %llu, size %zu", - (u64)bh->b_blocknr, bh->b_size); - } + if (!silent) + nilfs_msg(sb, KERN_WARNING, + "discard dirty block: blocknr=%llu, size=%zu", + (u64)bh->b_blocknr, bh->b_size); + set_mask_bits(&bh->b_state, clear_bits, 0); unlock_buffer(bh); } while (bh = bh->b_this_page, bh != head); diff --git a/fs/nilfs2/recovery.c b/fs/nilfs2/recovery.c index d893dc912b62..5139efed1888 100644 --- a/fs/nilfs2/recovery.c +++ b/fs/nilfs2/recovery.c @@ -54,38 +54,37 @@ struct nilfs_recovery_block { }; -static int nilfs_warn_segment_error(int err) +static int nilfs_warn_segment_error(struct super_block *sb, int err) { + const char *msg = NULL; + switch (err) { case NILFS_SEG_FAIL_IO: - printk(KERN_WARNING - "NILFS warning: I/O error on loading last segment\n"); + nilfs_msg(sb, KERN_ERR, "I/O error reading segment"); return -EIO; case NILFS_SEG_FAIL_MAGIC: - printk(KERN_WARNING - "NILFS warning: Segment magic number invalid\n"); + msg = "Magic number mismatch"; break; case NILFS_SEG_FAIL_SEQ: - printk(KERN_WARNING - "NILFS warning: Sequence number mismatch\n"); + msg = "Sequence number mismatch"; break; case NILFS_SEG_FAIL_CHECKSUM_SUPER_ROOT: - printk(KERN_WARNING - "NILFS warning: Checksum error in super root\n"); + msg = "Checksum error in super root"; break; case NILFS_SEG_FAIL_CHECKSUM_FULL: - printk(KERN_WARNING - "NILFS warning: Checksum error in segment payload\n"); + msg = "Checksum error in segment payload"; break; case NILFS_SEG_FAIL_CONSISTENCY: - printk(KERN_WARNING - "NILFS warning: Inconsistent segment\n"); + msg = "Inconsistency found"; break; case NILFS_SEG_NO_SUPER_ROOT: - printk(KERN_WARNING - "NILFS warning: No super root in the last segment\n"); + msg = "No super root in the last segment"; break; + default: + nilfs_msg(sb, KERN_ERR, "unrecognized segment error %d", err); + return -EINVAL; } + nilfs_msg(sb, KERN_WARNING, "invalid segment: %s", msg); return -EINVAL; } @@ -178,7 +177,7 @@ int nilfs_read_super_root_block(struct the_nilfs *nilfs, sector_t sr_block, brelse(bh_sr); failed: - return nilfs_warn_segment_error(ret); + return nilfs_warn_segment_error(nilfs->ns_sb, ret); } /** @@ -553,11 +552,10 @@ static int nilfs_recover_dsync_blocks(struct the_nilfs *nilfs, put_page(page); failed_inode: - printk(KERN_WARNING - "NILFS warning: error recovering data block " - "(err=%d, ino=%lu, block-offset=%llu)\n", - err, (unsigned long)rb->ino, - (unsigned long long)rb->blkoff); + nilfs_msg(sb, KERN_WARNING, + "error %d recovering data block (ino=%lu, block-offset=%llu)", + err, (unsigned long)rb->ino, + (unsigned long long)rb->blkoff); if (!err2) err2 = err; next: @@ -680,8 +678,8 @@ static int nilfs_do_roll_forward(struct the_nilfs *nilfs, } if (nsalvaged_blocks) { - printk(KERN_INFO "NILFS (device %s): salvaged %lu blocks\n", - sb->s_id, nsalvaged_blocks); + nilfs_msg(sb, KERN_INFO, "salvaged %lu blocks", + nsalvaged_blocks); ri->ri_need_recovery = NILFS_RECOVERY_ROLLFORWARD_DONE; } out: @@ -692,10 +690,9 @@ static int nilfs_do_roll_forward(struct the_nilfs *nilfs, confused: err = -EINVAL; failed: - printk(KERN_ERR - "NILFS (device %s): Error roll-forwarding " - "(err=%d, pseg block=%llu). ", - sb->s_id, err, (unsigned long long)pseg_start); + nilfs_msg(sb, KERN_ERR, + "error %d roll-forwarding partial segment at blocknr = %llu", + err, (unsigned long long)pseg_start); goto out; } @@ -715,9 +712,8 @@ static void nilfs_finish_roll_forward(struct the_nilfs *nilfs, set_buffer_dirty(bh); err = sync_dirty_buffer(bh); if (unlikely(err)) - printk(KERN_WARNING - "NILFS warning: buffer sync write failed during " - "post-cleaning of recovery.\n"); + nilfs_msg(nilfs->ns_sb, KERN_WARNING, + "buffer sync write failed during post-cleaning of recovery."); brelse(bh); } @@ -752,8 +748,8 @@ int nilfs_salvage_orphan_logs(struct the_nilfs *nilfs, err = nilfs_attach_checkpoint(sb, ri->ri_cno, true, &root); if (unlikely(err)) { - printk(KERN_ERR - "NILFS: error loading the latest checkpoint.\n"); + nilfs_msg(sb, KERN_ERR, + "error %d loading the latest checkpoint", err); return err; } @@ -764,8 +760,9 @@ int nilfs_salvage_orphan_logs(struct the_nilfs *nilfs, if (ri->ri_need_recovery == NILFS_RECOVERY_ROLLFORWARD_DONE) { err = nilfs_prepare_segment_for_recovery(nilfs, sb, ri); if (unlikely(err)) { - printk(KERN_ERR "NILFS: Error preparing segments for " - "recovery.\n"); + nilfs_msg(sb, KERN_ERR, + "error %d preparing segment for recovery", + err); goto failed; } @@ -778,8 +775,9 @@ int nilfs_salvage_orphan_logs(struct the_nilfs *nilfs, nilfs_detach_log_writer(sb); if (unlikely(err)) { - printk(KERN_ERR "NILFS: Oops! recovery failed. " - "(err=%d)\n", err); + nilfs_msg(sb, KERN_ERR, + "error %d writing segment for recovery", + err); goto failed; } @@ -961,5 +959,5 @@ int nilfs_search_super_root(struct the_nilfs *nilfs, failed: brelse(bh_sum); nilfs_dispose_segment_list(&segments); - return (ret < 0) ? ret : nilfs_warn_segment_error(ret); + return ret < 0 ? ret : nilfs_warn_segment_error(nilfs->ns_sb, ret); } diff --git a/fs/nilfs2/segbuf.c b/fs/nilfs2/segbuf.c index a962d7d83447..6f87b2ac1aeb 100644 --- a/fs/nilfs2/segbuf.c +++ b/fs/nilfs2/segbuf.c @@ -514,7 +514,11 @@ static int nilfs_segbuf_wait(struct nilfs_segment_buffer *segbuf) } while (--segbuf->sb_nbio > 0); if (unlikely(atomic_read(&segbuf->sb_err) > 0)) { - printk(KERN_ERR "NILFS: IO error writing segment\n"); + nilfs_msg(segbuf->sb_super, KERN_ERR, + "I/O error writing log (start-blocknr=%llu, block-count=%lu) in segment %llu", + (unsigned long long)segbuf->sb_pseg_start, + segbuf->sb_sum.nblocks, + (unsigned long long)segbuf->sb_segnum); err = -EIO; } return err; diff --git a/fs/nilfs2/segment.c b/fs/nilfs2/segment.c index e78b68a81aec..bedcae2c28e6 100644 --- a/fs/nilfs2/segment.c +++ b/fs/nilfs2/segment.c @@ -150,7 +150,8 @@ static void nilfs_dispose_list(struct the_nilfs *, struct list_head *, int); #define nilfs_cnt32_lt(a, b) nilfs_cnt32_gt(b, a) #define nilfs_cnt32_le(a, b) nilfs_cnt32_ge(b, a) -static int nilfs_prepare_segment_lock(struct nilfs_transaction_info *ti) +static int nilfs_prepare_segment_lock(struct super_block *sb, + struct nilfs_transaction_info *ti) { struct nilfs_transaction_info *cur_ti = current->journal_info; void *save = NULL; @@ -164,8 +165,7 @@ static int nilfs_prepare_segment_lock(struct nilfs_transaction_info *ti) * it is saved and will be restored on * nilfs_transaction_commit(). */ - printk(KERN_WARNING - "NILFS warning: journal info from a different FS\n"); + nilfs_msg(sb, KERN_WARNING, "journal info from a different FS"); save = current->journal_info; } if (!ti) { @@ -215,7 +215,7 @@ int nilfs_transaction_begin(struct super_block *sb, int vacancy_check) { struct the_nilfs *nilfs; - int ret = nilfs_prepare_segment_lock(ti); + int ret = nilfs_prepare_segment_lock(sb, ti); struct nilfs_transaction_info *trace_ti; if (unlikely(ret < 0)) @@ -373,7 +373,7 @@ static void nilfs_transaction_lock(struct super_block *sb, nilfs_segctor_do_immediate_flush(sci); up_write(&nilfs->ns_segctor_sem); - yield(); + cond_resched(); } if (gcflag) ti->ti_flags |= NILFS_TI_GC; @@ -1858,11 +1858,11 @@ static void nilfs_segctor_complete_write(struct nilfs_sc_info *sci) */ list_for_each_entry(bh, &segbuf->sb_payload_buffers, b_assoc_buffers) { - const unsigned long set_bits = (1 << BH_Uptodate); + const unsigned long set_bits = BIT(BH_Uptodate); const unsigned long clear_bits = - (1 << BH_Dirty | 1 << BH_Async_Write | - 1 << BH_Delay | 1 << BH_NILFS_Volatile | - 1 << BH_NILFS_Redirected); + (BIT(BH_Dirty) | BIT(BH_Async_Write) | + BIT(BH_Delay) | BIT(BH_NILFS_Volatile) | + BIT(BH_NILFS_Redirected)); set_mask_bits(&bh->b_state, clear_bits, set_bits); if (bh == segbuf->sb_super_root) { @@ -1951,8 +1951,9 @@ static int nilfs_segctor_collect_dirty_files(struct nilfs_sc_info *sci, err = nilfs_ifile_get_inode_block( ifile, ii->vfs_inode.i_ino, &ibh); if (unlikely(err)) { - nilfs_warning(sci->sc_super, __func__, - "failed to get inode block."); + nilfs_msg(sci->sc_super, KERN_WARNING, + "log writer: error %d getting inode block (ino=%lu)", + err, ii->vfs_inode.i_ino); return err; } mark_buffer_dirty(ibh); @@ -2131,10 +2132,10 @@ static void nilfs_segctor_start_timer(struct nilfs_sc_info *sci) static void nilfs_segctor_do_flush(struct nilfs_sc_info *sci, int bn) { spin_lock(&sci->sc_state_lock); - if (!(sci->sc_flush_request & (1 << bn))) { + if (!(sci->sc_flush_request & BIT(bn))) { unsigned long prev_req = sci->sc_flush_request; - sci->sc_flush_request |= (1 << bn); + sci->sc_flush_request |= BIT(bn); if (!prev_req) wake_up(&sci->sc_wait_daemon); } @@ -2318,7 +2319,7 @@ int nilfs_construct_dsync_segment(struct super_block *sb, struct inode *inode, } #define FLUSH_FILE_BIT (0x1) /* data file only */ -#define FLUSH_DAT_BIT (1 << NILFS_DAT_INO) /* DAT only */ +#define FLUSH_DAT_BIT BIT(NILFS_DAT_INO) /* DAT only */ /** * nilfs_segctor_accept - record accepted sequence count of log-write requests @@ -2458,8 +2459,7 @@ int nilfs_clean_segments(struct super_block *sb, struct nilfs_argv *argv, if (likely(!err)) break; - nilfs_warning(sb, __func__, - "segment construction failed. (err=%d)", err); + nilfs_msg(sb, KERN_WARNING, "error %d cleaning segments", err); set_current_state(TASK_INTERRUPTIBLE); schedule_timeout(sci->sc_interval); } @@ -2467,9 +2467,9 @@ int nilfs_clean_segments(struct super_block *sb, struct nilfs_argv *argv, int ret = nilfs_discard_segments(nilfs, sci->sc_freesegs, sci->sc_nfreesegs); if (ret) { - printk(KERN_WARNING - "NILFS warning: error %d on discard request, " - "turning discards off for the device\n", ret); + nilfs_msg(sb, KERN_WARNING, + "error %d on discard request, turning discards off for the device", + ret); nilfs_clear_opt(nilfs, DISCARD); } } @@ -2551,10 +2551,9 @@ static int nilfs_segctor_thread(void *arg) /* start sync. */ sci->sc_task = current; wake_up(&sci->sc_wait_task); /* for nilfs_segctor_start_thread() */ - printk(KERN_INFO - "segctord starting. Construction interval = %lu seconds, " - "CP frequency < %lu seconds\n", - sci->sc_interval / HZ, sci->sc_mjcp_freq / HZ); + nilfs_msg(sci->sc_super, KERN_INFO, + "segctord starting. Construction interval = %lu seconds, CP frequency < %lu seconds", + sci->sc_interval / HZ, sci->sc_mjcp_freq / HZ); spin_lock(&sci->sc_state_lock); loop: @@ -2628,8 +2627,8 @@ static int nilfs_segctor_start_thread(struct nilfs_sc_info *sci) if (IS_ERR(t)) { int err = PTR_ERR(t); - printk(KERN_ERR "NILFS: error %d creating segctord thread\n", - err); + nilfs_msg(sci->sc_super, KERN_ERR, + "error %d creating segctord thread", err); return err; } wait_event(sci->sc_wait_task, sci->sc_task != NULL); @@ -2739,14 +2738,14 @@ static void nilfs_segctor_destroy(struct nilfs_sc_info *sci) nilfs_segctor_write_out(sci); if (!list_empty(&sci->sc_dirty_files)) { - nilfs_warning(sci->sc_super, __func__, - "dirty file(s) after the final construction"); + nilfs_msg(sci->sc_super, KERN_WARNING, + "disposed unprocessed dirty file(s) when stopping log writer"); nilfs_dispose_list(nilfs, &sci->sc_dirty_files, 1); } if (!list_empty(&sci->sc_iput_queue)) { - nilfs_warning(sci->sc_super, __func__, - "iput queue is not empty"); + nilfs_msg(sci->sc_super, KERN_WARNING, + "disposed unprocessed inode(s) in iput queue when stopping log writer"); nilfs_dispose_list(nilfs, &sci->sc_iput_queue, 1); } @@ -2822,8 +2821,8 @@ void nilfs_detach_log_writer(struct super_block *sb) spin_lock(&nilfs->ns_inode_lock); if (!list_empty(&nilfs->ns_dirty_files)) { list_splice_init(&nilfs->ns_dirty_files, &garbage_list); - nilfs_warning(sb, __func__, - "Hit dirty file after stopped log writer"); + nilfs_msg(sb, KERN_WARNING, + "disposed unprocessed dirty file(s) when detaching log writer"); } spin_unlock(&nilfs->ns_inode_lock); up_write(&nilfs->ns_segctor_sem); diff --git a/fs/nilfs2/segment.h b/fs/nilfs2/segment.h index 6565c10b7b76..1060949d7dd2 100644 --- a/fs/nilfs2/segment.h +++ b/fs/nilfs2/segment.h @@ -23,7 +23,6 @@ #include <linux/fs.h> #include <linux/buffer_head.h> #include <linux/workqueue.h> -#include <linux/nilfs2_fs.h> #include "nilfs.h" struct nilfs_root; diff --git a/fs/nilfs2/sufile.c b/fs/nilfs2/sufile.c index 1963595a1580..1541a1e9221a 100644 --- a/fs/nilfs2/sufile.c +++ b/fs/nilfs2/sufile.c @@ -22,7 +22,6 @@ #include <linux/string.h> #include <linux/buffer_head.h> #include <linux/errno.h> -#include <linux/nilfs2_fs.h> #include "mdt.h" #include "sufile.h" @@ -181,9 +180,9 @@ int nilfs_sufile_updatev(struct inode *sufile, __u64 *segnumv, size_t nsegs, down_write(&NILFS_MDT(sufile)->mi_sem); for (seg = segnumv; seg < segnumv + nsegs; seg++) { if (unlikely(*seg >= nilfs_sufile_get_nsegments(sufile))) { - printk(KERN_WARNING - "%s: invalid segment number: %llu\n", __func__, - (unsigned long long)*seg); + nilfs_msg(sufile->i_sb, KERN_WARNING, + "%s: invalid segment number: %llu", + __func__, (unsigned long long)*seg); nerr++; } } @@ -240,8 +239,9 @@ int nilfs_sufile_update(struct inode *sufile, __u64 segnum, int create, int ret; if (unlikely(segnum >= nilfs_sufile_get_nsegments(sufile))) { - printk(KERN_WARNING "%s: invalid segment number: %llu\n", - __func__, (unsigned long long)segnum); + nilfs_msg(sufile->i_sb, KERN_WARNING, + "%s: invalid segment number: %llu", + __func__, (unsigned long long)segnum); return -EINVAL; } down_write(&NILFS_MDT(sufile)->mi_sem); @@ -419,8 +419,9 @@ void nilfs_sufile_do_cancel_free(struct inode *sufile, __u64 segnum, kaddr = kmap_atomic(su_bh->b_page); su = nilfs_sufile_block_get_segment_usage(sufile, segnum, su_bh, kaddr); if (unlikely(!nilfs_segment_usage_clean(su))) { - printk(KERN_WARNING "%s: segment %llu must be clean\n", - __func__, (unsigned long long)segnum); + nilfs_msg(sufile->i_sb, KERN_WARNING, + "%s: segment %llu must be clean", __func__, + (unsigned long long)segnum); kunmap_atomic(kaddr); return; } @@ -444,7 +445,7 @@ void nilfs_sufile_do_scrap(struct inode *sufile, __u64 segnum, kaddr = kmap_atomic(su_bh->b_page); su = nilfs_sufile_block_get_segment_usage(sufile, segnum, su_bh, kaddr); - if (su->su_flags == cpu_to_le32(1UL << NILFS_SEGMENT_USAGE_DIRTY) && + if (su->su_flags == cpu_to_le32(BIT(NILFS_SEGMENT_USAGE_DIRTY)) && su->su_nblocks == cpu_to_le32(0)) { kunmap_atomic(kaddr); return; @@ -455,7 +456,7 @@ void nilfs_sufile_do_scrap(struct inode *sufile, __u64 segnum, /* make the segment garbage */ su->su_lastmod = cpu_to_le64(0); su->su_nblocks = cpu_to_le32(0); - su->su_flags = cpu_to_le32(1UL << NILFS_SEGMENT_USAGE_DIRTY); + su->su_flags = cpu_to_le32(BIT(NILFS_SEGMENT_USAGE_DIRTY)); kunmap_atomic(kaddr); nilfs_sufile_mod_counter(header_bh, clean ? (u64)-1 : 0, dirty ? 0 : 1); @@ -476,8 +477,9 @@ void nilfs_sufile_do_free(struct inode *sufile, __u64 segnum, kaddr = kmap_atomic(su_bh->b_page); su = nilfs_sufile_block_get_segment_usage(sufile, segnum, su_bh, kaddr); if (nilfs_segment_usage_clean(su)) { - printk(KERN_WARNING "%s: segment %llu is already clean\n", - __func__, (unsigned long long)segnum); + nilfs_msg(sufile->i_sb, KERN_WARNING, + "%s: segment %llu is already clean", + __func__, (unsigned long long)segnum); kunmap_atomic(kaddr); return; } @@ -692,7 +694,7 @@ static int nilfs_sufile_truncate_range(struct inode *sufile, su2 = su; for (j = 0; j < n; j++, su = (void *)su + susz) { if ((le32_to_cpu(su->su_flags) & - ~(1UL << NILFS_SEGMENT_USAGE_ERROR)) || + ~BIT(NILFS_SEGMENT_USAGE_ERROR)) || nilfs_segment_is_active(nilfs, segnum + j)) { ret = -EBUSY; kunmap_atomic(kaddr); @@ -859,10 +861,10 @@ ssize_t nilfs_sufile_get_suinfo(struct inode *sufile, __u64 segnum, void *buf, si->sui_lastmod = le64_to_cpu(su->su_lastmod); si->sui_nblocks = le32_to_cpu(su->su_nblocks); si->sui_flags = le32_to_cpu(su->su_flags) & - ~(1UL << NILFS_SEGMENT_USAGE_ACTIVE); + ~BIT(NILFS_SEGMENT_USAGE_ACTIVE); if (nilfs_segment_is_active(nilfs, segnum + j)) si->sui_flags |= - (1UL << NILFS_SEGMENT_USAGE_ACTIVE); + BIT(NILFS_SEGMENT_USAGE_ACTIVE); } kunmap_atomic(kaddr); brelse(su_bh); @@ -950,7 +952,7 @@ ssize_t nilfs_sufile_set_suinfo(struct inode *sufile, void *buf, * disk. */ sup->sup_sui.sui_flags &= - ~(1UL << NILFS_SEGMENT_USAGE_ACTIVE); + ~BIT(NILFS_SEGMENT_USAGE_ACTIVE); cleansi = nilfs_suinfo_clean(&sup->sup_sui); cleansu = nilfs_segment_usage_clean(su); @@ -1175,14 +1177,12 @@ int nilfs_sufile_read(struct super_block *sb, size_t susize, int err; if (susize > sb->s_blocksize) { - printk(KERN_ERR - "NILFS: too large segment usage size: %zu bytes.\n", - susize); + nilfs_msg(sb, KERN_ERR, + "too large segment usage size: %zu bytes", susize); return -EINVAL; } else if (susize < NILFS_MIN_SEGMENT_USAGE_SIZE) { - printk(KERN_ERR - "NILFS: too small segment usage size: %zu bytes.\n", - susize); + nilfs_msg(sb, KERN_ERR, + "too small segment usage size: %zu bytes", susize); return -EINVAL; } diff --git a/fs/nilfs2/sufile.h b/fs/nilfs2/sufile.h index 46e89872294c..158a9190c8ec 100644 --- a/fs/nilfs2/sufile.h +++ b/fs/nilfs2/sufile.h @@ -21,7 +21,6 @@ #include <linux/fs.h> #include <linux/buffer_head.h> -#include <linux/nilfs2_fs.h> #include "mdt.h" diff --git a/fs/nilfs2/super.c b/fs/nilfs2/super.c index 666107a18a22..c95d369e90aa 100644 --- a/fs/nilfs2/super.c +++ b/fs/nilfs2/super.c @@ -71,6 +71,22 @@ struct kmem_cache *nilfs_btree_path_cache; static int nilfs_setup_super(struct super_block *sb, int is_mount); static int nilfs_remount(struct super_block *sb, int *flags, char *data); +void __nilfs_msg(struct super_block *sb, const char *level, const char *fmt, + ...) +{ + struct va_format vaf; + va_list args; + + va_start(args, fmt); + vaf.fmt = fmt; + vaf.va = &args; + if (sb) + printk("%sNILFS (%s): %pV\n", level, sb->s_id, &vaf); + else + printk("%sNILFS: %pV\n", level, &vaf); + va_end(args); +} + static void nilfs_set_error(struct super_block *sb) { struct the_nilfs *nilfs = sb->s_fs_info; @@ -91,19 +107,20 @@ static void nilfs_set_error(struct super_block *sb) } /** - * nilfs_error() - report failure condition on a filesystem + * __nilfs_error() - report failure condition on a filesystem + * + * __nilfs_error() sets an ERROR_FS flag on the superblock as well as + * reporting an error message. This function should be called when + * NILFS detects incoherences or defects of meta data on disk. * - * nilfs_error() sets an ERROR_FS flag on the superblock as well as - * reporting an error message. It should be called when NILFS detects - * incoherences or defects of meta data on disk. As for sustainable - * errors such as a single-shot I/O error, nilfs_warning() or the printk() - * function should be used instead. + * This implements the body of nilfs_error() macro. Normally, + * nilfs_error() should be used. As for sustainable errors such as a + * single-shot I/O error, nilfs_msg() should be used instead. * - * The segment constructor must not call this function because it can - * kill itself. + * Callers should not add a trailing newline since this will do it. */ -void nilfs_error(struct super_block *sb, const char *function, - const char *fmt, ...) +void __nilfs_error(struct super_block *sb, const char *function, + const char *fmt, ...) { struct the_nilfs *nilfs = sb->s_fs_info; struct va_format vaf; @@ -133,24 +150,6 @@ void nilfs_error(struct super_block *sb, const char *function, sb->s_id); } -void nilfs_warning(struct super_block *sb, const char *function, - const char *fmt, ...) -{ - struct va_format vaf; - va_list args; - - va_start(args, fmt); - - vaf.fmt = fmt; - vaf.va = &args; - - printk(KERN_WARNING "NILFS warning (device %s): %s: %pV\n", - sb->s_id, function, &vaf); - - va_end(args); -} - - struct inode *nilfs_alloc_inode(struct super_block *sb) { struct nilfs_inode_info *ii; @@ -196,8 +195,8 @@ static int nilfs_sync_super(struct super_block *sb, int flag) } if (unlikely(err)) { - printk(KERN_ERR - "NILFS: unable to write superblock (err=%d)\n", err); + nilfs_msg(sb, KERN_ERR, "unable to write superblock: err=%d", + err); if (err == -EIO && nilfs->ns_sbh[1]) { /* * sbp[0] points to newer log than sbp[1], @@ -267,8 +266,7 @@ struct nilfs_super_block **nilfs_prepare_super(struct super_block *sb, sbp[1]->s_magic == cpu_to_le16(NILFS_SUPER_MAGIC)) { memcpy(sbp[0], sbp[1], nilfs->ns_sbsize); } else { - printk(KERN_CRIT "NILFS: superblock broke on dev %s\n", - sb->s_id); + nilfs_msg(sb, KERN_CRIT, "superblock broke"); return NULL; } } else if (sbp[1] && @@ -378,9 +376,9 @@ static int nilfs_move_2nd_super(struct super_block *sb, loff_t sb2off) offset = sb2off & (nilfs->ns_blocksize - 1); nsbh = sb_getblk(sb, newblocknr); if (!nsbh) { - printk(KERN_WARNING - "NILFS warning: unable to move secondary superblock " - "to block %llu\n", (unsigned long long)newblocknr); + nilfs_msg(sb, KERN_WARNING, + "unable to move secondary superblock to block %llu", + (unsigned long long)newblocknr); ret = -EIO; goto out; } @@ -543,10 +541,9 @@ int nilfs_attach_checkpoint(struct super_block *sb, __u64 cno, int curr_mnt, up_read(&nilfs->ns_segctor_sem); if (unlikely(err)) { if (err == -ENOENT || err == -EINVAL) { - printk(KERN_ERR - "NILFS: Invalid checkpoint " - "(checkpoint number=%llu)\n", - (unsigned long long)cno); + nilfs_msg(sb, KERN_ERR, + "Invalid checkpoint (checkpoint number=%llu)", + (unsigned long long)cno); err = -EINVAL; } goto failed; @@ -642,9 +639,8 @@ static int nilfs_statfs(struct dentry *dentry, struct kstatfs *buf) err = nilfs_ifile_count_free_inodes(root->ifile, &nmaxinodes, &nfreeinodes); if (unlikely(err)) { - printk(KERN_WARNING - "NILFS warning: fail to count free inodes: err %d.\n", - err); + nilfs_msg(sb, KERN_WARNING, + "failed to count free inodes: err=%d", err); if (err == -ERANGE) { /* * If nilfs_palloc_count_max_entries() returns @@ -776,9 +772,9 @@ static int parse_options(char *options, struct super_block *sb, int is_remount) break; case Opt_snapshot: if (is_remount) { - printk(KERN_ERR - "NILFS: \"%s\" option is invalid " - "for remount.\n", p); + nilfs_msg(sb, KERN_ERR, + "\"%s\" option is invalid for remount", + p); return 0; } break; @@ -792,8 +788,8 @@ static int parse_options(char *options, struct super_block *sb, int is_remount) nilfs_clear_opt(nilfs, DISCARD); break; default: - printk(KERN_ERR - "NILFS: Unrecognized mount option \"%s\"\n", p); + nilfs_msg(sb, KERN_ERR, + "unrecognized mount option \"%s\"", p); return 0; } } @@ -829,12 +825,10 @@ static int nilfs_setup_super(struct super_block *sb, int is_mount) mnt_count = le16_to_cpu(sbp[0]->s_mnt_count); if (nilfs->ns_mount_state & NILFS_ERROR_FS) { - printk(KERN_WARNING - "NILFS warning: mounting fs with errors\n"); + nilfs_msg(sb, KERN_WARNING, "mounting fs with errors"); #if 0 } else if (max_mnt_count >= 0 && mnt_count >= max_mnt_count) { - printk(KERN_WARNING - "NILFS warning: maximal mount count reached\n"); + nilfs_msg(sb, KERN_WARNING, "maximal mount count reached"); #endif } if (!max_mnt_count) @@ -897,17 +891,17 @@ int nilfs_check_feature_compatibility(struct super_block *sb, features = le64_to_cpu(sbp->s_feature_incompat) & ~NILFS_FEATURE_INCOMPAT_SUPP; if (features) { - printk(KERN_ERR "NILFS: couldn't mount because of unsupported " - "optional features (%llx)\n", - (unsigned long long)features); + nilfs_msg(sb, KERN_ERR, + "couldn't mount because of unsupported optional features (%llx)", + (unsigned long long)features); return -EINVAL; } features = le64_to_cpu(sbp->s_feature_compat_ro) & ~NILFS_FEATURE_COMPAT_RO_SUPP; if (!(sb->s_flags & MS_RDONLY) && features) { - printk(KERN_ERR "NILFS: couldn't mount RDWR because of " - "unsupported optional features (%llx)\n", - (unsigned long long)features); + nilfs_msg(sb, KERN_ERR, + "couldn't mount RDWR because of unsupported optional features (%llx)", + (unsigned long long)features); return -EINVAL; } return 0; @@ -923,13 +917,13 @@ static int nilfs_get_root_dentry(struct super_block *sb, inode = nilfs_iget(sb, root, NILFS_ROOT_INO); if (IS_ERR(inode)) { - printk(KERN_ERR "NILFS: get root inode failed\n"); ret = PTR_ERR(inode); + nilfs_msg(sb, KERN_ERR, "error %d getting root inode", ret); goto out; } if (!S_ISDIR(inode->i_mode) || !inode->i_blocks || !inode->i_size) { iput(inode); - printk(KERN_ERR "NILFS: corrupt root inode.\n"); + nilfs_msg(sb, KERN_ERR, "corrupt root inode"); ret = -EINVAL; goto out; } @@ -957,7 +951,7 @@ static int nilfs_get_root_dentry(struct super_block *sb, return ret; failed_dentry: - printk(KERN_ERR "NILFS: get root dentry failed\n"); + nilfs_msg(sb, KERN_ERR, "error %d getting root dentry", ret); goto out; } @@ -977,18 +971,18 @@ static int nilfs_attach_snapshot(struct super_block *s, __u64 cno, ret = (ret == -ENOENT) ? -EINVAL : ret; goto out; } else if (!ret) { - printk(KERN_ERR "NILFS: The specified checkpoint is " - "not a snapshot (checkpoint number=%llu).\n", - (unsigned long long)cno); + nilfs_msg(s, KERN_ERR, + "The specified checkpoint is not a snapshot (checkpoint number=%llu)", + (unsigned long long)cno); ret = -EINVAL; goto out; } ret = nilfs_attach_checkpoint(s, cno, false, &root); if (ret) { - printk(KERN_ERR "NILFS: error loading snapshot " - "(checkpoint number=%llu).\n", - (unsigned long long)cno); + nilfs_msg(s, KERN_ERR, + "error %d while loading snapshot (checkpoint number=%llu)", + ret, (unsigned long long)cno); goto out; } ret = nilfs_get_root_dentry(s, root, root_dentry); @@ -1058,7 +1052,7 @@ nilfs_fill_super(struct super_block *sb, void *data, int silent) __u64 cno; int err; - nilfs = alloc_nilfs(sb->s_bdev); + nilfs = alloc_nilfs(sb); if (!nilfs) return -ENOMEM; @@ -1083,8 +1077,9 @@ nilfs_fill_super(struct super_block *sb, void *data, int silent) cno = nilfs_last_cno(nilfs); err = nilfs_attach_checkpoint(sb, cno, true, &fsroot); if (err) { - printk(KERN_ERR "NILFS: error loading last checkpoint " - "(checkpoint number=%llu).\n", (unsigned long long)cno); + nilfs_msg(sb, KERN_ERR, + "error %d while loading last checkpoint (checkpoint number=%llu)", + err, (unsigned long long)cno); goto failed_unload; } @@ -1144,9 +1139,8 @@ static int nilfs_remount(struct super_block *sb, int *flags, char *data) err = -EINVAL; if (!nilfs_valid_fs(nilfs)) { - printk(KERN_WARNING "NILFS (device %s): couldn't " - "remount because the filesystem is in an " - "incomplete recovery state.\n", sb->s_id); + nilfs_msg(sb, KERN_WARNING, + "couldn't remount because the filesystem is in an incomplete recovery state"); goto restore_opts; } @@ -1178,10 +1172,9 @@ static int nilfs_remount(struct super_block *sb, int *flags, char *data) ~NILFS_FEATURE_COMPAT_RO_SUPP; up_read(&nilfs->ns_sem); if (features) { - printk(KERN_WARNING "NILFS (device %s): couldn't " - "remount RDWR because of unsupported optional " - "features (%llx)\n", - sb->s_id, (unsigned long long)features); + nilfs_msg(sb, KERN_WARNING, + "couldn't remount RDWR because of unsupported optional features (%llx)", + (unsigned long long)features); err = -EROFS; goto restore_opts; } @@ -1212,6 +1205,38 @@ struct nilfs_super_data { int flags; }; +static int nilfs_parse_snapshot_option(const char *option, + const substring_t *arg, + struct nilfs_super_data *sd) +{ + unsigned long long val; + const char *msg = NULL; + int err; + + if (!(sd->flags & MS_RDONLY)) { + msg = "read-only option is not specified"; + goto parse_error; + } + + err = kstrtoull(arg->from, 0, &val); + if (err) { + if (err == -ERANGE) + msg = "too large checkpoint number"; + else + msg = "malformed argument"; + goto parse_error; + } else if (val == 0) { + msg = "invalid checkpoint number 0"; + goto parse_error; + } + sd->cno = val; + return 0; + +parse_error: + nilfs_msg(NULL, KERN_ERR, "invalid option \"%s\": %s", option, msg); + return 1; +} + /** * nilfs_identify - pre-read mount options needed to identify mount instance * @data: mount options @@ -1228,24 +1253,9 @@ static int nilfs_identify(char *data, struct nilfs_super_data *sd) p = strsep(&options, ","); if (p != NULL && *p) { token = match_token(p, tokens, args); - if (token == Opt_snapshot) { - if (!(sd->flags & MS_RDONLY)) { - ret++; - } else { - sd->cno = simple_strtoull(args[0].from, - NULL, 0); - /* - * No need to see the end pointer; - * match_token() has done syntax - * checking. - */ - if (sd->cno == 0) - ret++; - } - } - if (ret) - printk(KERN_ERR - "NILFS: invalid mount option: %s\n", p); + if (token == Opt_snapshot) + ret = nilfs_parse_snapshot_option(p, &args[0], + sd); } if (!options) break; @@ -1326,10 +1336,10 @@ nilfs_mount(struct file_system_type *fs_type, int flags, } else if (!sd.cno) { if (nilfs_tree_is_busy(s->s_root)) { if ((flags ^ s->s_flags) & MS_RDONLY) { - printk(KERN_ERR "NILFS: the device already " - "has a %s mount.\n", - (s->s_flags & MS_RDONLY) ? - "read-only" : "read/write"); + nilfs_msg(s, KERN_ERR, + "the device already has a %s mount.", + (s->s_flags & MS_RDONLY) ? + "read-only" : "read/write"); err = -EBUSY; goto failed_super; } diff --git a/fs/nilfs2/sysfs.c b/fs/nilfs2/sysfs.c index 8ffa42b704d8..490303e3d517 100644 --- a/fs/nilfs2/sysfs.c +++ b/fs/nilfs2/sysfs.c @@ -272,8 +272,8 @@ nilfs_checkpoints_checkpoints_number_show(struct nilfs_checkpoints_attr *attr, err = nilfs_cpfile_get_stat(nilfs->ns_cpfile, &cpstat); up_read(&nilfs->ns_segctor_sem); if (err < 0) { - printk(KERN_ERR "NILFS: unable to get checkpoint stat: err=%d\n", - err); + nilfs_msg(nilfs->ns_sb, KERN_ERR, + "unable to get checkpoint stat: err=%d", err); return err; } @@ -295,8 +295,8 @@ nilfs_checkpoints_snapshots_number_show(struct nilfs_checkpoints_attr *attr, err = nilfs_cpfile_get_stat(nilfs->ns_cpfile, &cpstat); up_read(&nilfs->ns_segctor_sem); if (err < 0) { - printk(KERN_ERR "NILFS: unable to get checkpoint stat: err=%d\n", - err); + nilfs_msg(nilfs->ns_sb, KERN_ERR, + "unable to get checkpoint stat: err=%d", err); return err; } @@ -326,9 +326,9 @@ nilfs_checkpoints_next_checkpoint_show(struct nilfs_checkpoints_attr *attr, { __u64 cno; - down_read(&nilfs->ns_sem); + down_read(&nilfs->ns_segctor_sem); cno = nilfs->ns_cno; - up_read(&nilfs->ns_sem); + up_read(&nilfs->ns_segctor_sem); return snprintf(buf, PAGE_SIZE, "%llu\n", cno); } @@ -414,8 +414,8 @@ nilfs_segments_dirty_segments_show(struct nilfs_segments_attr *attr, err = nilfs_sufile_get_stat(nilfs->ns_sufile, &sustat); up_read(&nilfs->ns_segctor_sem); if (err < 0) { - printk(KERN_ERR "NILFS: unable to get segment stat: err=%d\n", - err); + nilfs_msg(nilfs->ns_sb, KERN_ERR, + "unable to get segment stat: err=%d", err); return err; } @@ -511,9 +511,9 @@ nilfs_segctor_current_seg_sequence_show(struct nilfs_segctor_attr *attr, { u64 seg_seq; - down_read(&nilfs->ns_sem); + down_read(&nilfs->ns_segctor_sem); seg_seq = nilfs->ns_seg_seq; - up_read(&nilfs->ns_sem); + up_read(&nilfs->ns_segctor_sem); return snprintf(buf, PAGE_SIZE, "%llu\n", seg_seq); } @@ -525,9 +525,9 @@ nilfs_segctor_current_last_full_seg_show(struct nilfs_segctor_attr *attr, { __u64 segnum; - down_read(&nilfs->ns_sem); + down_read(&nilfs->ns_segctor_sem); segnum = nilfs->ns_segnum; - up_read(&nilfs->ns_sem); + up_read(&nilfs->ns_segctor_sem); return snprintf(buf, PAGE_SIZE, "%llu\n", segnum); } @@ -539,9 +539,9 @@ nilfs_segctor_next_full_seg_show(struct nilfs_segctor_attr *attr, { __u64 nextnum; - down_read(&nilfs->ns_sem); + down_read(&nilfs->ns_segctor_sem); nextnum = nilfs->ns_nextnum; - up_read(&nilfs->ns_sem); + up_read(&nilfs->ns_segctor_sem); return snprintf(buf, PAGE_SIZE, "%llu\n", nextnum); } @@ -553,9 +553,9 @@ nilfs_segctor_next_pseg_offset_show(struct nilfs_segctor_attr *attr, { unsigned long pseg_offset; - down_read(&nilfs->ns_sem); + down_read(&nilfs->ns_segctor_sem); pseg_offset = nilfs->ns_pseg_offset; - up_read(&nilfs->ns_sem); + up_read(&nilfs->ns_segctor_sem); return snprintf(buf, PAGE_SIZE, "%lu\n", pseg_offset); } @@ -567,9 +567,9 @@ nilfs_segctor_next_checkpoint_show(struct nilfs_segctor_attr *attr, { __u64 cno; - down_read(&nilfs->ns_sem); + down_read(&nilfs->ns_segctor_sem); cno = nilfs->ns_cno; - up_read(&nilfs->ns_sem); + up_read(&nilfs->ns_segctor_sem); return snprintf(buf, PAGE_SIZE, "%llu\n", cno); } @@ -581,9 +581,9 @@ nilfs_segctor_last_seg_write_time_show(struct nilfs_segctor_attr *attr, { time_t ctime; - down_read(&nilfs->ns_sem); + down_read(&nilfs->ns_segctor_sem); ctime = nilfs->ns_ctime; - up_read(&nilfs->ns_sem); + up_read(&nilfs->ns_segctor_sem); return NILFS_SHOW_TIME(ctime, buf); } @@ -595,9 +595,9 @@ nilfs_segctor_last_seg_write_time_secs_show(struct nilfs_segctor_attr *attr, { time_t ctime; - down_read(&nilfs->ns_sem); + down_read(&nilfs->ns_segctor_sem); ctime = nilfs->ns_ctime; - up_read(&nilfs->ns_sem); + up_read(&nilfs->ns_segctor_sem); return snprintf(buf, PAGE_SIZE, "%llu\n", (unsigned long long)ctime); } @@ -609,9 +609,9 @@ nilfs_segctor_last_nongc_write_time_show(struct nilfs_segctor_attr *attr, { time_t nongc_ctime; - down_read(&nilfs->ns_sem); + down_read(&nilfs->ns_segctor_sem); nongc_ctime = nilfs->ns_nongc_ctime; - up_read(&nilfs->ns_sem); + up_read(&nilfs->ns_segctor_sem); return NILFS_SHOW_TIME(nongc_ctime, buf); } @@ -623,9 +623,9 @@ nilfs_segctor_last_nongc_write_time_secs_show(struct nilfs_segctor_attr *attr, { time_t nongc_ctime; - down_read(&nilfs->ns_sem); + down_read(&nilfs->ns_segctor_sem); nongc_ctime = nilfs->ns_nongc_ctime; - up_read(&nilfs->ns_sem); + up_read(&nilfs->ns_segctor_sem); return snprintf(buf, PAGE_SIZE, "%llu\n", (unsigned long long)nongc_ctime); @@ -638,9 +638,9 @@ nilfs_segctor_dirty_data_blocks_count_show(struct nilfs_segctor_attr *attr, { u32 ndirtyblks; - down_read(&nilfs->ns_sem); + down_read(&nilfs->ns_segctor_sem); ndirtyblks = atomic_read(&nilfs->ns_ndirtyblks); - up_read(&nilfs->ns_sem); + up_read(&nilfs->ns_segctor_sem); return snprintf(buf, PAGE_SIZE, "%u\n", ndirtyblks); } @@ -789,14 +789,15 @@ nilfs_superblock_sb_update_frequency_store(struct nilfs_superblock_attr *attr, err = kstrtouint(skip_spaces(buf), 0, &val); if (err) { - printk(KERN_ERR "NILFS: unable to convert string: err=%d\n", - err); + nilfs_msg(nilfs->ns_sb, KERN_ERR, + "unable to convert string: err=%d", err); return err; } if (val < NILFS_SB_FREQ) { val = NILFS_SB_FREQ; - printk(KERN_WARNING "NILFS: superblock update frequency cannot be lesser than 10 seconds\n"); + nilfs_msg(nilfs->ns_sb, KERN_WARNING, + "superblock update frequency cannot be lesser than 10 seconds"); } down_write(&nilfs->ns_sem); @@ -999,7 +1000,8 @@ int nilfs_sysfs_create_device_group(struct super_block *sb) nilfs->ns_dev_subgroups = kzalloc(devgrp_size, GFP_KERNEL); if (unlikely(!nilfs->ns_dev_subgroups)) { err = -ENOMEM; - printk(KERN_ERR "NILFS: unable to allocate memory for device group\n"); + nilfs_msg(sb, KERN_ERR, + "unable to allocate memory for device group"); goto failed_create_device_group; } @@ -1109,15 +1111,15 @@ int __init nilfs_sysfs_init(void) nilfs_kset = kset_create_and_add(NILFS_ROOT_GROUP_NAME, NULL, fs_kobj); if (!nilfs_kset) { err = -ENOMEM; - printk(KERN_ERR "NILFS: unable to create sysfs entry: err %d\n", - err); + nilfs_msg(NULL, KERN_ERR, + "unable to create sysfs entry: err=%d", err); goto failed_sysfs_init; } err = sysfs_create_group(&nilfs_kset->kobj, &nilfs_feature_attr_group); if (unlikely(err)) { - printk(KERN_ERR "NILFS: unable to create feature group: err %d\n", - err); + nilfs_msg(NULL, KERN_ERR, + "unable to create feature group: err=%d", err); goto cleanup_sysfs_init; } diff --git a/fs/nilfs2/the_nilfs.c b/fs/nilfs2/the_nilfs.c index e9fd241b9a0a..2dd75bf619ad 100644 --- a/fs/nilfs2/the_nilfs.c +++ b/fs/nilfs2/the_nilfs.c @@ -56,12 +56,12 @@ void nilfs_set_last_segment(struct the_nilfs *nilfs, /** * alloc_nilfs - allocate a nilfs object - * @bdev: block device to which the_nilfs is related + * @sb: super block instance * * Return Value: On success, pointer to the_nilfs is returned. * On error, NULL is returned. */ -struct the_nilfs *alloc_nilfs(struct block_device *bdev) +struct the_nilfs *alloc_nilfs(struct super_block *sb) { struct the_nilfs *nilfs; @@ -69,7 +69,8 @@ struct the_nilfs *alloc_nilfs(struct block_device *bdev) if (!nilfs) return NULL; - nilfs->ns_bdev = bdev; + nilfs->ns_sb = sb; + nilfs->ns_bdev = sb->s_bdev; atomic_set(&nilfs->ns_ndirtyblks, 0); init_rwsem(&nilfs->ns_sem); mutex_init(&nilfs->ns_snapshot_mount_mutex); @@ -191,7 +192,10 @@ static int nilfs_store_log_cursor(struct the_nilfs *nilfs, nilfs_get_segnum_of_block(nilfs, nilfs->ns_last_pseg); nilfs->ns_cno = nilfs->ns_last_cno + 1; if (nilfs->ns_segnum >= nilfs->ns_nsegments) { - printk(KERN_ERR "NILFS invalid last segment number.\n"); + nilfs_msg(nilfs->ns_sb, KERN_ERR, + "pointed segment number is out of range: segnum=%llu, nsegments=%lu", + (unsigned long long)nilfs->ns_segnum, + nilfs->ns_nsegments); ret = -EINVAL; } return ret; @@ -215,12 +219,12 @@ int load_nilfs(struct the_nilfs *nilfs, struct super_block *sb) int err; if (!valid_fs) { - printk(KERN_WARNING "NILFS warning: mounting unchecked fs\n"); + nilfs_msg(sb, KERN_WARNING, "mounting unchecked fs"); if (s_flags & MS_RDONLY) { - printk(KERN_INFO "NILFS: INFO: recovery " - "required for readonly filesystem.\n"); - printk(KERN_INFO "NILFS: write access will " - "be enabled during recovery.\n"); + nilfs_msg(sb, KERN_INFO, + "recovery required for readonly filesystem"); + nilfs_msg(sb, KERN_INFO, + "write access will be enabled during recovery"); } } @@ -235,13 +239,12 @@ int load_nilfs(struct the_nilfs *nilfs, struct super_block *sb) goto scan_error; if (!nilfs_valid_sb(sbp[1])) { - printk(KERN_WARNING - "NILFS warning: unable to fall back to spare" - "super block\n"); + nilfs_msg(sb, KERN_WARNING, + "unable to fall back to spare super block"); goto scan_error; } - printk(KERN_INFO - "NILFS: try rollback from an earlier position\n"); + nilfs_msg(sb, KERN_INFO, + "trying rollback from an earlier position"); /* * restore super block with its spare and reconfigure @@ -254,10 +257,9 @@ int load_nilfs(struct the_nilfs *nilfs, struct super_block *sb) /* verify consistency between two super blocks */ blocksize = BLOCK_SIZE << le32_to_cpu(sbp[0]->s_log_block_size); if (blocksize != nilfs->ns_blocksize) { - printk(KERN_WARNING - "NILFS warning: blocksize differs between " - "two super blocks (%d != %d)\n", - blocksize, nilfs->ns_blocksize); + nilfs_msg(sb, KERN_WARNING, + "blocksize differs between two super blocks (%d != %d)", + blocksize, nilfs->ns_blocksize); goto scan_error; } @@ -276,7 +278,8 @@ int load_nilfs(struct the_nilfs *nilfs, struct super_block *sb) err = nilfs_load_super_root(nilfs, sb, ri.ri_super_root); if (unlikely(err)) { - printk(KERN_ERR "NILFS: error loading super root.\n"); + nilfs_msg(sb, KERN_ERR, "error %d while loading super root", + err); goto failed; } @@ -287,30 +290,29 @@ int load_nilfs(struct the_nilfs *nilfs, struct super_block *sb) __u64 features; if (nilfs_test_opt(nilfs, NORECOVERY)) { - printk(KERN_INFO "NILFS: norecovery option specified. " - "skipping roll-forward recovery\n"); + nilfs_msg(sb, KERN_INFO, + "norecovery option specified, skipping roll-forward recovery"); goto skip_recovery; } features = le64_to_cpu(nilfs->ns_sbp[0]->s_feature_compat_ro) & ~NILFS_FEATURE_COMPAT_RO_SUPP; if (features) { - printk(KERN_ERR "NILFS: couldn't proceed with " - "recovery because of unsupported optional " - "features (%llx)\n", - (unsigned long long)features); + nilfs_msg(sb, KERN_ERR, + "couldn't proceed with recovery because of unsupported optional features (%llx)", + (unsigned long long)features); err = -EROFS; goto failed_unload; } if (really_read_only) { - printk(KERN_ERR "NILFS: write access " - "unavailable, cannot proceed.\n"); + nilfs_msg(sb, KERN_ERR, + "write access unavailable, cannot proceed"); err = -EROFS; goto failed_unload; } sb->s_flags &= ~MS_RDONLY; } else if (nilfs_test_opt(nilfs, NORECOVERY)) { - printk(KERN_ERR "NILFS: recovery cancelled because norecovery " - "option was specified for a read/write mount\n"); + nilfs_msg(sb, KERN_ERR, + "recovery cancelled because norecovery option was specified for a read/write mount"); err = -EINVAL; goto failed_unload; } @@ -325,11 +327,12 @@ int load_nilfs(struct the_nilfs *nilfs, struct super_block *sb) up_write(&nilfs->ns_sem); if (err) { - printk(KERN_ERR "NILFS: failed to update super block. " - "recovery unfinished.\n"); + nilfs_msg(sb, KERN_ERR, + "error %d updating super block. recovery unfinished.", + err); goto failed_unload; } - printk(KERN_INFO "NILFS: recovery complete.\n"); + nilfs_msg(sb, KERN_INFO, "recovery complete"); skip_recovery: nilfs_clear_recovery_info(&ri); @@ -337,7 +340,7 @@ int load_nilfs(struct the_nilfs *nilfs, struct super_block *sb) return 0; scan_error: - printk(KERN_ERR "NILFS: error searching super root.\n"); + nilfs_msg(sb, KERN_ERR, "error %d while searching super root", err); goto failed; failed_unload: @@ -384,12 +387,11 @@ static int nilfs_store_disk_layout(struct the_nilfs *nilfs, struct nilfs_super_block *sbp) { if (le32_to_cpu(sbp->s_rev_level) < NILFS_MIN_SUPP_REV) { - printk(KERN_ERR "NILFS: unsupported revision " - "(superblock rev.=%d.%d, current rev.=%d.%d). " - "Please check the version of mkfs.nilfs.\n", - le32_to_cpu(sbp->s_rev_level), - le16_to_cpu(sbp->s_minor_rev_level), - NILFS_CURRENT_REV, NILFS_MINOR_REV); + nilfs_msg(nilfs->ns_sb, KERN_ERR, + "unsupported revision (superblock rev.=%d.%d, current rev.=%d.%d). Please check the version of mkfs.nilfs(2).", + le32_to_cpu(sbp->s_rev_level), + le16_to_cpu(sbp->s_minor_rev_level), + NILFS_CURRENT_REV, NILFS_MINOR_REV); return -EINVAL; } nilfs->ns_sbsize = le16_to_cpu(sbp->s_bytes); @@ -398,12 +400,14 @@ static int nilfs_store_disk_layout(struct the_nilfs *nilfs, nilfs->ns_inode_size = le16_to_cpu(sbp->s_inode_size); if (nilfs->ns_inode_size > nilfs->ns_blocksize) { - printk(KERN_ERR "NILFS: too large inode size: %d bytes.\n", - nilfs->ns_inode_size); + nilfs_msg(nilfs->ns_sb, KERN_ERR, + "too large inode size: %d bytes", + nilfs->ns_inode_size); return -EINVAL; } else if (nilfs->ns_inode_size < NILFS_MIN_INODE_SIZE) { - printk(KERN_ERR "NILFS: too small inode size: %d bytes.\n", - nilfs->ns_inode_size); + nilfs_msg(nilfs->ns_sb, KERN_ERR, + "too small inode size: %d bytes", + nilfs->ns_inode_size); return -EINVAL; } @@ -411,7 +415,9 @@ static int nilfs_store_disk_layout(struct the_nilfs *nilfs, nilfs->ns_blocks_per_segment = le32_to_cpu(sbp->s_blocks_per_segment); if (nilfs->ns_blocks_per_segment < NILFS_SEG_MIN_BLOCKS) { - printk(KERN_ERR "NILFS: too short segment.\n"); + nilfs_msg(nilfs->ns_sb, KERN_ERR, + "too short segment: %lu blocks", + nilfs->ns_blocks_per_segment); return -EINVAL; } @@ -420,7 +426,9 @@ static int nilfs_store_disk_layout(struct the_nilfs *nilfs, le32_to_cpu(sbp->s_r_segments_percentage); if (nilfs->ns_r_segments_percentage < 1 || nilfs->ns_r_segments_percentage > 99) { - printk(KERN_ERR "NILFS: invalid reserved segments percentage.\n"); + nilfs_msg(nilfs->ns_sb, KERN_ERR, + "invalid reserved segments percentage: %lu", + nilfs->ns_r_segments_percentage); return -EINVAL; } @@ -504,16 +512,16 @@ static int nilfs_load_super_block(struct the_nilfs *nilfs, if (!sbp[0]) { if (!sbp[1]) { - printk(KERN_ERR "NILFS: unable to read superblock\n"); + nilfs_msg(sb, KERN_ERR, "unable to read superblock"); return -EIO; } - printk(KERN_WARNING - "NILFS warning: unable to read primary superblock " - "(blocksize = %d)\n", blocksize); + nilfs_msg(sb, KERN_WARNING, + "unable to read primary superblock (blocksize = %d)", + blocksize); } else if (!sbp[1]) { - printk(KERN_WARNING - "NILFS warning: unable to read secondary superblock " - "(blocksize = %d)\n", blocksize); + nilfs_msg(sb, KERN_WARNING, + "unable to read secondary superblock (blocksize = %d)", + blocksize); } /* @@ -535,14 +543,14 @@ static int nilfs_load_super_block(struct the_nilfs *nilfs, } if (!valid[swp]) { nilfs_release_super_block(nilfs); - printk(KERN_ERR "NILFS: Can't find nilfs on dev %s.\n", - sb->s_id); + nilfs_msg(sb, KERN_ERR, "couldn't find nilfs on the device"); return -EINVAL; } if (!valid[!swp]) - printk(KERN_WARNING "NILFS warning: broken superblock. " - "using spare superblock (blocksize = %d).\n", blocksize); + nilfs_msg(sb, KERN_WARNING, + "broken superblock, retrying with spare superblock (blocksize = %d)", + blocksize); if (swp) nilfs_swap_super_block(nilfs); @@ -576,7 +584,7 @@ int init_nilfs(struct the_nilfs *nilfs, struct super_block *sb, char *data) blocksize = sb_min_blocksize(sb, NILFS_MIN_BLOCK_SIZE); if (!blocksize) { - printk(KERN_ERR "NILFS: unable to set blocksize\n"); + nilfs_msg(sb, KERN_ERR, "unable to set blocksize"); err = -EINVAL; goto out; } @@ -595,8 +603,9 @@ int init_nilfs(struct the_nilfs *nilfs, struct super_block *sb, char *data) blocksize = BLOCK_SIZE << le32_to_cpu(sbp->s_log_block_size); if (blocksize < NILFS_MIN_BLOCK_SIZE || blocksize > NILFS_MAX_BLOCK_SIZE) { - printk(KERN_ERR "NILFS: couldn't mount because of unsupported " - "filesystem blocksize %d\n", blocksize); + nilfs_msg(sb, KERN_ERR, + "couldn't mount because of unsupported filesystem blocksize %d", + blocksize); err = -EINVAL; goto failed_sbh; } @@ -604,10 +613,9 @@ int init_nilfs(struct the_nilfs *nilfs, struct super_block *sb, char *data) int hw_blocksize = bdev_logical_block_size(sb->s_bdev); if (blocksize < hw_blocksize) { - printk(KERN_ERR - "NILFS: blocksize %d too small for device " - "(sector-size = %d).\n", - blocksize, hw_blocksize); + nilfs_msg(sb, KERN_ERR, + "blocksize %d too small for device (sector-size = %d)", + blocksize, hw_blocksize); err = -EINVAL; goto failed_sbh; } diff --git a/fs/nilfs2/the_nilfs.h b/fs/nilfs2/the_nilfs.h index 79369fd6b13b..b305c6f033e7 100644 --- a/fs/nilfs2/the_nilfs.h +++ b/fs/nilfs2/the_nilfs.h @@ -43,6 +43,7 @@ enum { * struct the_nilfs - struct to supervise multiple nilfs mount points * @ns_flags: flags * @ns_flushed_device: flag indicating if all volatile data was flushed + * @ns_sb: back pointer to super block instance * @ns_bdev: block device * @ns_sem: semaphore for shared states * @ns_snapshot_mount_mutex: mutex to protect snapshot mounts @@ -102,6 +103,7 @@ struct the_nilfs { unsigned long ns_flags; int ns_flushed_device; + struct super_block *ns_sb; struct block_device *ns_bdev; struct rw_semaphore ns_sem; struct mutex ns_snapshot_mount_mutex; @@ -120,11 +122,8 @@ struct the_nilfs { unsigned int ns_sb_update_freq; /* - * Following fields are dedicated to a writable FS-instance. - * Except for the period seeking checkpoint, code outside the segment - * constructor must lock a segment semaphore while accessing these - * fields. - * The writable FS-instance is sole during a lifetime of the_nilfs. + * The following fields are updated by a writable FS-instance. + * These fields are protected by ns_segctor_sem outside load_nilfs(). */ u64 ns_seg_seq; __u64 ns_segnum; @@ -281,7 +280,7 @@ static inline int nilfs_sb_will_flip(struct the_nilfs *nilfs) } void nilfs_set_last_segment(struct the_nilfs *, sector_t, u64, __u64); -struct the_nilfs *alloc_nilfs(struct block_device *bdev); +struct the_nilfs *alloc_nilfs(struct super_block *sb); void destroy_nilfs(struct the_nilfs *nilfs); int init_nilfs(struct the_nilfs *nilfs, struct super_block *sb, char *data); int load_nilfs(struct the_nilfs *nilfs, struct super_block *sb); diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c index 460c0cedab3a..7dabbc31060e 100644 --- a/fs/ocfs2/alloc.c +++ b/fs/ocfs2/alloc.c @@ -6106,6 +6106,43 @@ void ocfs2_schedule_truncate_log_flush(struct ocfs2_super *osb, } } +/* + * Try to flush truncate logs if we can free enough clusters from it. + * As for return value, "< 0" means error, "0" no space and "1" means + * we have freed enough spaces and let the caller try to allocate again. + */ +int ocfs2_try_to_free_truncate_log(struct ocfs2_super *osb, + unsigned int needed) +{ + tid_t target; + int ret = 0; + unsigned int truncated_clusters; + + inode_lock(osb->osb_tl_inode); + truncated_clusters = osb->truncated_clusters; + inode_unlock(osb->osb_tl_inode); + + /* + * Check whether we can succeed in allocating if we free + * the truncate log. + */ + if (truncated_clusters < needed) + goto out; + + ret = ocfs2_flush_truncate_log(osb); + if (ret) { + mlog_errno(ret); + goto out; + } + + if (jbd2_journal_start_commit(osb->journal->j_journal, &target)) { + jbd2_log_wait_commit(osb->journal->j_journal, target); + ret = 1; + } +out: + return ret; +} + static int ocfs2_get_truncate_log_info(struct ocfs2_super *osb, int slot_num, struct inode **tl_inode, diff --git a/fs/ocfs2/alloc.h b/fs/ocfs2/alloc.h index f3dc1b0dfffc..4a5152ec88a3 100644 --- a/fs/ocfs2/alloc.h +++ b/fs/ocfs2/alloc.h @@ -188,6 +188,8 @@ int ocfs2_truncate_log_append(struct ocfs2_super *osb, u64 start_blk, unsigned int num_clusters); int __ocfs2_flush_truncate_log(struct ocfs2_super *osb); +int ocfs2_try_to_free_truncate_log(struct ocfs2_super *osb, + unsigned int needed); /* * Process local structure which describes the block unlinks done diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c index af2adfcb0f6f..98d36548153d 100644 --- a/fs/ocfs2/aops.c +++ b/fs/ocfs2/aops.c @@ -1645,43 +1645,6 @@ static int ocfs2_zero_tail(struct inode *inode, struct buffer_head *di_bh, return ret; } -/* - * Try to flush truncate logs if we can free enough clusters from it. - * As for return value, "< 0" means error, "0" no space and "1" means - * we have freed enough spaces and let the caller try to allocate again. - */ -static int ocfs2_try_to_free_truncate_log(struct ocfs2_super *osb, - unsigned int needed) -{ - tid_t target; - int ret = 0; - unsigned int truncated_clusters; - - inode_lock(osb->osb_tl_inode); - truncated_clusters = osb->truncated_clusters; - inode_unlock(osb->osb_tl_inode); - - /* - * Check whether we can succeed in allocating if we free - * the truncate log. - */ - if (truncated_clusters < needed) - goto out; - - ret = ocfs2_flush_truncate_log(osb); - if (ret) { - mlog_errno(ret); - goto out; - } - - if (jbd2_journal_start_commit(osb->journal->j_journal, &target)) { - jbd2_log_wait_commit(osb->journal->j_journal, target); - ret = 1; - } -out: - return ret; -} - int ocfs2_write_begin_nolock(struct address_space *mapping, loff_t pos, unsigned len, ocfs2_write_type_t type, struct page **pagep, void **fsdata, diff --git a/fs/ocfs2/dlm/dlmcommon.h b/fs/ocfs2/dlm/dlmcommon.h index 8107d0d0c3f6..e9f3705c4c9f 100644 --- a/fs/ocfs2/dlm/dlmcommon.h +++ b/fs/ocfs2/dlm/dlmcommon.h @@ -1004,6 +1004,8 @@ int dlm_finalize_reco_handler(struct o2net_msg *msg, u32 len, void *data, int dlm_do_master_requery(struct dlm_ctxt *dlm, struct dlm_lock_resource *res, u8 nodenum, u8 *real_master); +void __dlm_do_purge_lockres(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res); int dlm_dispatch_assert_master(struct dlm_ctxt *dlm, struct dlm_lock_resource *res, diff --git a/fs/ocfs2/dlm/dlmmaster.c b/fs/ocfs2/dlm/dlmmaster.c index 13719d3f35f8..6ea06f8a7d29 100644 --- a/fs/ocfs2/dlm/dlmmaster.c +++ b/fs/ocfs2/dlm/dlmmaster.c @@ -2276,9 +2276,12 @@ int dlm_drop_lockres_ref(struct dlm_ctxt *dlm, struct dlm_lock_resource *res) mlog(ML_ERROR, "%s: res %.*s, DEREF to node %u got %d\n", dlm->name, namelen, lockname, res->owner, r); dlm_print_one_lock_resource(res); - BUG(); - } - return ret ? ret : r; + if (r == -ENOMEM) + BUG(); + } else + ret = r; + + return ret; } int dlm_deref_lockres_handler(struct o2net_msg *msg, u32 len, void *data, @@ -2416,48 +2419,26 @@ int dlm_deref_lockres_done_handler(struct o2net_msg *msg, u32 len, void *data, } spin_lock(&res->spinlock); - BUG_ON(!(res->state & DLM_LOCK_RES_DROPPING_REF)); - if (!list_empty(&res->purge)) { - mlog(0, "%s: Removing res %.*s from purgelist\n", - dlm->name, res->lockname.len, res->lockname.name); - list_del_init(&res->purge); - dlm_lockres_put(res); - dlm->purge_count--; - } - - if (!__dlm_lockres_unused(res)) { - mlog(ML_ERROR, "%s: res %.*s in use after deref\n", - dlm->name, res->lockname.len, res->lockname.name); - __dlm_print_one_lock_resource(res); - BUG(); - } - - __dlm_unhash_lockres(dlm, res); - - spin_lock(&dlm->track_lock); - if (!list_empty(&res->tracking)) - list_del_init(&res->tracking); - else { - mlog(ML_ERROR, "%s: Resource %.*s not on the Tracking list\n", - dlm->name, res->lockname.len, res->lockname.name); - __dlm_print_one_lock_resource(res); + if (!(res->state & DLM_LOCK_RES_DROPPING_REF)) { + spin_unlock(&res->spinlock); + spin_unlock(&dlm->spinlock); + mlog(ML_NOTICE, "%s:%.*s: node %u sends deref done " + "but it is already derefed!\n", dlm->name, + res->lockname.len, res->lockname.name, node); + ret = 0; + goto done; } - spin_unlock(&dlm->track_lock); - /* lockres is not in the hash now. drop the flag and wake up - * any processes waiting in dlm_get_lock_resource. - */ - res->state &= ~DLM_LOCK_RES_DROPPING_REF; + __dlm_do_purge_lockres(dlm, res); spin_unlock(&res->spinlock); wake_up(&res->wq); - dlm_lockres_put(res); - spin_unlock(&dlm->spinlock); ret = 0; - done: + if (res) + dlm_lockres_put(res); dlm_put(dlm); return ret; } diff --git a/fs/ocfs2/dlm/dlmrecovery.c b/fs/ocfs2/dlm/dlmrecovery.c index f6b313898763..dd5cb8bcefd1 100644 --- a/fs/ocfs2/dlm/dlmrecovery.c +++ b/fs/ocfs2/dlm/dlmrecovery.c @@ -2343,6 +2343,7 @@ static void dlm_do_local_recovery_cleanup(struct dlm_ctxt *dlm, u8 dead_node) struct dlm_lock_resource *res; int i; struct hlist_head *bucket; + struct hlist_node *tmp; struct dlm_lock *lock; @@ -2365,7 +2366,7 @@ static void dlm_do_local_recovery_cleanup(struct dlm_ctxt *dlm, u8 dead_node) */ for (i = 0; i < DLM_HASH_BUCKETS; i++) { bucket = dlm_lockres_hash(dlm, i); - hlist_for_each_entry(res, bucket, hash_node) { + hlist_for_each_entry_safe(res, tmp, bucket, hash_node) { /* always prune any $RECOVERY entries for dead nodes, * otherwise hangs can occur during later recovery */ if (dlm_is_recovery_lock(res->lockname.name, @@ -2386,8 +2387,17 @@ static void dlm_do_local_recovery_cleanup(struct dlm_ctxt *dlm, u8 dead_node) break; } } - dlm_lockres_clear_refmap_bit(dlm, res, - dead_node); + + if ((res->owner == dead_node) && + (res->state & DLM_LOCK_RES_DROPPING_REF)) { + dlm_lockres_get(res); + __dlm_do_purge_lockres(dlm, res); + spin_unlock(&res->spinlock); + wake_up(&res->wq); + dlm_lockres_put(res); + continue; + } else if (res->owner == dlm->node_num) + dlm_lockres_clear_refmap_bit(dlm, res, dead_node); spin_unlock(&res->spinlock); continue; } @@ -2398,14 +2408,17 @@ static void dlm_do_local_recovery_cleanup(struct dlm_ctxt *dlm, u8 dead_node) if (res->state & DLM_LOCK_RES_DROPPING_REF) { mlog(0, "%s:%.*s: owned by " "dead node %u, this node was " - "dropping its ref when it died. " - "continue, dropping the flag.\n", + "dropping its ref when master died. " + "continue, purging the lockres.\n", dlm->name, res->lockname.len, res->lockname.name, dead_node); + dlm_lockres_get(res); + __dlm_do_purge_lockres(dlm, res); + spin_unlock(&res->spinlock); + wake_up(&res->wq); + dlm_lockres_put(res); + continue; } - res->state &= ~DLM_LOCK_RES_DROPPING_REF; - dlm_move_lockres_to_recovery_list(dlm, - res); } else if (res->owner == dlm->node_num) { dlm_free_dead_locks(dlm, res, dead_node); __dlm_lockres_calc_usage(dlm, res); diff --git a/fs/ocfs2/dlm/dlmthread.c b/fs/ocfs2/dlm/dlmthread.c index 68d239ba0c63..838a06d4066a 100644 --- a/fs/ocfs2/dlm/dlmthread.c +++ b/fs/ocfs2/dlm/dlmthread.c @@ -160,6 +160,52 @@ void dlm_lockres_calc_usage(struct dlm_ctxt *dlm, spin_unlock(&dlm->spinlock); } +/* + * Do the real purge work: + * unhash the lockres, and + * clear flag DLM_LOCK_RES_DROPPING_REF. + * It requires dlm and lockres spinlock to be taken. + */ +void __dlm_do_purge_lockres(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res) +{ + assert_spin_locked(&dlm->spinlock); + assert_spin_locked(&res->spinlock); + + if (!list_empty(&res->purge)) { + mlog(0, "%s: Removing res %.*s from purgelist\n", + dlm->name, res->lockname.len, res->lockname.name); + list_del_init(&res->purge); + dlm_lockres_put(res); + dlm->purge_count--; + } + + if (!__dlm_lockres_unused(res)) { + mlog(ML_ERROR, "%s: res %.*s in use after deref\n", + dlm->name, res->lockname.len, res->lockname.name); + __dlm_print_one_lock_resource(res); + BUG(); + } + + __dlm_unhash_lockres(dlm, res); + + spin_lock(&dlm->track_lock); + if (!list_empty(&res->tracking)) + list_del_init(&res->tracking); + else { + mlog(ML_ERROR, "%s: Resource %.*s not on the Tracking list\n", + dlm->name, res->lockname.len, res->lockname.name); + __dlm_print_one_lock_resource(res); + } + spin_unlock(&dlm->track_lock); + + /* + * lockres is not in the hash now. drop the flag and wake up + * any processes waiting in dlm_get_lock_resource. + */ + res->state &= ~DLM_LOCK_RES_DROPPING_REF; +} + static void dlm_purge_lockres(struct dlm_ctxt *dlm, struct dlm_lock_resource *res) { @@ -175,6 +221,13 @@ static void dlm_purge_lockres(struct dlm_ctxt *dlm, res->lockname.len, res->lockname.name, master); if (!master) { + if (res->state & DLM_LOCK_RES_DROPPING_REF) { + mlog(ML_NOTICE, "%s: res %.*s already in DLM_LOCK_RES_DROPPING_REF state\n", + dlm->name, res->lockname.len, res->lockname.name); + spin_unlock(&res->spinlock); + return; + } + res->state |= DLM_LOCK_RES_DROPPING_REF; /* drop spinlock... retake below */ spin_unlock(&res->spinlock); @@ -203,8 +256,8 @@ static void dlm_purge_lockres(struct dlm_ctxt *dlm, dlm->purge_count--; } - if (!master && ret != 0) { - mlog(0, "%s: deref %.*s in progress or master goes down\n", + if (!master && ret == DLM_DEREF_RESPONSE_INPROG) { + mlog(0, "%s: deref %.*s in progress\n", dlm->name, res->lockname.len, res->lockname.name); spin_unlock(&res->spinlock); return; diff --git a/fs/ocfs2/stack_user.c b/fs/ocfs2/stack_user.c index ced70c8139f7..c9e828ec3c8e 100644 --- a/fs/ocfs2/stack_user.c +++ b/fs/ocfs2/stack_user.c @@ -1007,10 +1007,17 @@ static int user_cluster_connect(struct ocfs2_cluster_connection *conn) lc->oc_type = NO_CONTROLD; rc = dlm_new_lockspace(conn->cc_name, conn->cc_cluster_name, - DLM_LSFL_FS, DLM_LVB_LEN, + DLM_LSFL_FS | DLM_LSFL_NEWEXCL, DLM_LVB_LEN, &ocfs2_ls_ops, conn, &ops_rv, &fsdlm); - if (rc) + if (rc) { + if (rc == -EEXIST || rc == -EPROTO) + printk(KERN_ERR "ocfs2: Unable to create the " + "lockspace %s (%d), because a ocfs2-tools " + "program is running on this file system " + "with the same name lockspace\n", + conn->cc_name, rc); goto out; + } if (ops_rv == -EOPNOTSUPP) { lc->oc_type = WITH_CONTROLD; diff --git a/fs/ocfs2/suballoc.c b/fs/ocfs2/suballoc.c index 2f19aeec5482..ea47120a85ff 100644 --- a/fs/ocfs2/suballoc.c +++ b/fs/ocfs2/suballoc.c @@ -1164,7 +1164,8 @@ static int ocfs2_reserve_clusters_with_limit(struct ocfs2_super *osb, int flags, struct ocfs2_alloc_context **ac) { - int status; + int status, ret = 0; + int retried = 0; *ac = kzalloc(sizeof(struct ocfs2_alloc_context), GFP_KERNEL); if (!(*ac)) { @@ -1189,7 +1190,24 @@ static int ocfs2_reserve_clusters_with_limit(struct ocfs2_super *osb, } if (status == -ENOSPC) { +retry: status = ocfs2_reserve_cluster_bitmap_bits(osb, *ac); + /* Retry if there is sufficient space cached in truncate log */ + if (status == -ENOSPC && !retried) { + retried = 1; + ocfs2_inode_unlock((*ac)->ac_inode, 1); + inode_unlock((*ac)->ac_inode); + + ret = ocfs2_try_to_free_truncate_log(osb, bits_wanted); + if (ret == 1) + goto retry; + + if (ret < 0) + mlog_errno(ret); + + inode_lock((*ac)->ac_inode); + ocfs2_inode_lock((*ac)->ac_inode, NULL, 1); + } if (status < 0) { if (status != -ENOSPC) mlog_errno(status); diff --git a/fs/orangefs/dcache.c b/fs/orangefs/dcache.c index 5dfc4f3cfe68..00235bf644dc 100644 --- a/fs/orangefs/dcache.c +++ b/fs/orangefs/dcache.c @@ -73,6 +73,7 @@ static int orangefs_revalidate_lookup(struct dentry *dentry) } } + dentry->d_time = jiffies + dcache_timeout_msecs*HZ/1000; ret = 1; out_release_op: op_release(new_op); @@ -94,6 +95,9 @@ static int orangefs_d_revalidate(struct dentry *dentry, unsigned int flags) { int ret; + if (time_before(jiffies, dentry->d_time)) + return 1; + if (flags & LOOKUP_RCU) return -ECHILD; diff --git a/fs/orangefs/inode.c b/fs/orangefs/inode.c index 2e63e6d0a68e..28a0557a69be 100644 --- a/fs/orangefs/inode.c +++ b/fs/orangefs/inode.c @@ -262,7 +262,7 @@ int orangefs_getattr(struct vfsmount *mnt, "orangefs_getattr: called on %s\n", dentry->d_name.name); - ret = orangefs_inode_getattr(inode, 0, 1); + ret = orangefs_inode_getattr(inode, 0, 0); if (ret == 0) { generic_fillattr(inode, kstat); @@ -384,7 +384,7 @@ struct inode *orangefs_iget(struct super_block *sb, struct orangefs_object_kref if (!inode || !(inode->i_state & I_NEW)) return inode; - error = orangefs_inode_getattr(inode, 1, 0); + error = orangefs_inode_getattr(inode, 1, 1); if (error) { iget_failed(inode); return ERR_PTR(error); @@ -429,7 +429,7 @@ struct inode *orangefs_new_inode(struct super_block *sb, struct inode *dir, orangefs_set_inode(inode, ref); inode->i_ino = hash; /* needed for stat etc */ - error = orangefs_inode_getattr(inode, 1, 0); + error = orangefs_inode_getattr(inode, 1, 1); if (error) goto out_iput; diff --git a/fs/orangefs/namei.c b/fs/orangefs/namei.c index 7e8dfa97c44a..62c525936ee8 100644 --- a/fs/orangefs/namei.c +++ b/fs/orangefs/namei.c @@ -72,6 +72,8 @@ static int orangefs_create(struct inode *dir, d_instantiate(dentry, inode); unlock_new_inode(inode); + dentry->d_time = jiffies + dcache_timeout_msecs*HZ/1000; + ORANGEFS_I(inode)->getattr_time = jiffies - 1; gossip_debug(GOSSIP_NAME_DEBUG, "%s: dentry instantiated for %s\n", @@ -181,6 +183,8 @@ static struct dentry *orangefs_lookup(struct inode *dir, struct dentry *dentry, goto out; } + dentry->d_time = jiffies + dcache_timeout_msecs*HZ/1000; + inode = orangefs_iget(dir->i_sb, &new_op->downcall.resp.lookup.refn); if (IS_ERR(inode)) { gossip_debug(GOSSIP_NAME_DEBUG, @@ -189,6 +193,8 @@ static struct dentry *orangefs_lookup(struct inode *dir, struct dentry *dentry, goto out; } + ORANGEFS_I(inode)->getattr_time = jiffies - 1; + gossip_debug(GOSSIP_NAME_DEBUG, "%s:%s:%d " "Found good inode [%lu] with count [%d]\n", @@ -316,6 +322,8 @@ static int orangefs_symlink(struct inode *dir, d_instantiate(dentry, inode); unlock_new_inode(inode); + dentry->d_time = jiffies + dcache_timeout_msecs*HZ/1000; + ORANGEFS_I(inode)->getattr_time = jiffies - 1; gossip_debug(GOSSIP_NAME_DEBUG, "Inode (Symlink) %pU -> %s\n", @@ -378,6 +386,8 @@ static int orangefs_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode d_instantiate(dentry, inode); unlock_new_inode(inode); + dentry->d_time = jiffies + dcache_timeout_msecs*HZ/1000; + ORANGEFS_I(inode)->getattr_time = jiffies - 1; gossip_debug(GOSSIP_NAME_DEBUG, "Inode (Directory) %pU -> %s\n", @@ -408,6 +418,8 @@ static int orangefs_rename(struct inode *old_dir, "orangefs_rename: called (%pd2 => %pd2) ct=%d\n", old_dentry, new_dentry, d_count(new_dentry)); + ORANGEFS_I(new_dentry->d_parent->d_inode)->getattr_time = jiffies - 1; + new_op = op_alloc(ORANGEFS_VFS_OP_RENAME); if (!new_op) return -EINVAL; diff --git a/fs/orangefs/orangefs-kernel.h b/fs/orangefs/orangefs-kernel.h index 4b6e132d5a0f..633c07a6e3d8 100644 --- a/fs/orangefs/orangefs-kernel.h +++ b/fs/orangefs/orangefs-kernel.h @@ -246,6 +246,8 @@ struct orangefs_inode_s { * with this object */ unsigned long pinode_flags; + + unsigned long getattr_time; }; #define P_ATIME_FLAG 0 @@ -527,7 +529,7 @@ int orangefs_inode_setxattr(struct inode *inode, size_t size, int flags); -int orangefs_inode_getattr(struct inode *inode, int new, int size); +int orangefs_inode_getattr(struct inode *inode, int new, int bypass); int orangefs_inode_check_changed(struct inode *inode); @@ -546,6 +548,8 @@ extern struct mutex request_mutex; extern int debug; extern int op_timeout_secs; extern int slot_timeout_secs; +extern int dcache_timeout_msecs; +extern int getattr_timeout_msecs; extern struct list_head orangefs_superblocks; extern spinlock_t orangefs_superblocks_lock; extern struct list_head orangefs_request_list; diff --git a/fs/orangefs/orangefs-mod.c b/fs/orangefs/orangefs-mod.c index 6f072a8c0de1..e9fd5755c05f 100644 --- a/fs/orangefs/orangefs-mod.c +++ b/fs/orangefs/orangefs-mod.c @@ -47,6 +47,8 @@ struct client_debug_mask client_debug_mask = { NULL, 0, 0 }; unsigned int kernel_mask_set_mod_init; /* implicitly false */ int op_timeout_secs = ORANGEFS_DEFAULT_OP_TIMEOUT_SECS; int slot_timeout_secs = ORANGEFS_DEFAULT_SLOT_TIMEOUT_SECS; +int dcache_timeout_msecs = 50; +int getattr_timeout_msecs = 50; MODULE_LICENSE("GPL"); MODULE_AUTHOR("ORANGEFS Development Team"); diff --git a/fs/orangefs/orangefs-sysfs.c b/fs/orangefs/orangefs-sysfs.c index 5c03113e3ad2..375708c2db87 100644 --- a/fs/orangefs/orangefs-sysfs.c +++ b/fs/orangefs/orangefs-sysfs.c @@ -61,10 +61,21 @@ * Slots are requested and waited for, * the wait times out after slot_timeout_secs. * + * What: /sys/fs/orangefs/dcache_timeout_msecs + * Date: Jul 2016 + * Contact: Martin Brandenburg <martin@omnibond.com> + * Description: + * Time lookup is valid in milliseconds. + * + * What: /sys/fs/orangefs/getattr_timeout_msecs + * Date: Jul 2016 + * Contact: Martin Brandenburg <martin@omnibond.com> + * Description: + * Time getattr is valid in milliseconds. * * What: /sys/fs/orangefs/acache/... * Date: Jun 2015 - * Contact: Mike Marshall <hubcap@omnibond.com> + * Contact: Martin Brandenburg <martin@omnibond.com> * Description: * Attribute cache configurable settings. * @@ -117,6 +128,8 @@ struct orangefs_obj { int perf_history_size; int perf_time_interval_secs; int slot_timeout_secs; + int dcache_timeout_msecs; + int getattr_timeout_msecs; }; struct acache_orangefs_obj { @@ -658,6 +671,20 @@ static ssize_t sysfs_int_show(char *kobj_id, char *buf, void *attr) "%d\n", slot_timeout_secs); goto out; + } else if (!strcmp(orangefs_attr->attr.name, + "dcache_timeout_msecs")) { + rc = scnprintf(buf, + PAGE_SIZE, + "%d\n", + dcache_timeout_msecs); + goto out; + } else if (!strcmp(orangefs_attr->attr.name, + "getattr_timeout_msecs")) { + rc = scnprintf(buf, + PAGE_SIZE, + "%d\n", + getattr_timeout_msecs); + goto out; } else { goto out; } @@ -734,6 +761,12 @@ static ssize_t int_store(struct orangefs_obj *orangefs_obj, } else if (!strcmp(attr->attr.name, "slot_timeout_secs")) { rc = kstrtoint(buf, 0, &slot_timeout_secs); goto out; + } else if (!strcmp(attr->attr.name, "dcache_timeout_msecs")) { + rc = kstrtoint(buf, 0, &dcache_timeout_msecs); + goto out; + } else if (!strcmp(attr->attr.name, "getattr_timeout_msecs")) { + rc = kstrtoint(buf, 0, &getattr_timeout_msecs); + goto out; } else { goto out; } @@ -1361,6 +1394,12 @@ static struct orangefs_attribute op_timeout_secs_attribute = static struct orangefs_attribute slot_timeout_secs_attribute = __ATTR(slot_timeout_secs, 0664, int_orangefs_show, int_store); +static struct orangefs_attribute dcache_timeout_msecs_attribute = + __ATTR(dcache_timeout_msecs, 0664, int_orangefs_show, int_store); + +static struct orangefs_attribute getattr_timeout_msecs_attribute = + __ATTR(getattr_timeout_msecs, 0664, int_orangefs_show, int_store); + static struct orangefs_attribute perf_counter_reset_attribute = __ATTR(perf_counter_reset, 0664, @@ -1382,6 +1421,8 @@ static struct orangefs_attribute perf_time_interval_secs_attribute = static struct attribute *orangefs_default_attrs[] = { &op_timeout_secs_attribute.attr, &slot_timeout_secs_attribute.attr, + &dcache_timeout_msecs_attribute.attr, + &getattr_timeout_msecs_attribute.attr, &perf_counter_reset_attribute.attr, &perf_history_size_attribute.attr, &perf_time_interval_secs_attribute.attr, diff --git a/fs/orangefs/orangefs-utils.c b/fs/orangefs/orangefs-utils.c index c5fbc62357c6..d13c7291fd05 100644 --- a/fs/orangefs/orangefs-utils.c +++ b/fs/orangefs/orangefs-utils.c @@ -251,7 +251,7 @@ static int orangefs_inode_is_stale(struct inode *inode, int new, return 0; } -int orangefs_inode_getattr(struct inode *inode, int new, int size) +int orangefs_inode_getattr(struct inode *inode, int new, int bypass) { struct orangefs_inode_s *orangefs_inode = ORANGEFS_I(inode); struct orangefs_kernel_op_s *new_op; @@ -261,12 +261,16 @@ int orangefs_inode_getattr(struct inode *inode, int new, int size) gossip_debug(GOSSIP_UTILS_DEBUG, "%s: called on inode %pU\n", __func__, get_khandle_from_ino(inode)); + if (!new && !bypass) { + if (time_before(jiffies, orangefs_inode->getattr_time)) + return 0; + } + new_op = op_alloc(ORANGEFS_VFS_OP_GETATTR); if (!new_op) return -ENOMEM; new_op->upcall.req.getattr.refn = orangefs_inode->refn; - new_op->upcall.req.getattr.mask = size ? - ORANGEFS_ATTR_SYS_ALL_NOHINT : ORANGEFS_ATTR_SYS_ALL_NOHINT_NOSIZE; + new_op->upcall.req.getattr.mask = ORANGEFS_ATTR_SYS_ALL_NOHINT; ret = service_operation(new_op, __func__, get_interruptible_flag(inode)); @@ -287,20 +291,18 @@ int orangefs_inode_getattr(struct inode *inode, int new, int size) case S_IFREG: inode->i_flags = orangefs_inode_flags(&new_op-> downcall.resp.getattr.attributes); - if (size) { - inode_size = (loff_t)new_op-> - downcall.resp.getattr.attributes.size; - rounded_up_size = - (inode_size + (4096 - (inode_size % 4096))); - inode->i_size = inode_size; - orangefs_inode->blksize = - new_op->downcall.resp.getattr.attributes.blksize; - spin_lock(&inode->i_lock); - inode->i_bytes = inode_size; - inode->i_blocks = - (unsigned long)(rounded_up_size / 512); - spin_unlock(&inode->i_lock); - } + inode_size = (loff_t)new_op-> + downcall.resp.getattr.attributes.size; + rounded_up_size = + (inode_size + (4096 - (inode_size % 4096))); + inode->i_size = inode_size; + orangefs_inode->blksize = + new_op->downcall.resp.getattr.attributes.blksize; + spin_lock(&inode->i_lock); + inode->i_bytes = inode_size; + inode->i_blocks = + (unsigned long)(rounded_up_size / 512); + spin_unlock(&inode->i_lock); break; case S_IFDIR: inode->i_size = PAGE_SIZE; @@ -345,6 +347,7 @@ int orangefs_inode_getattr(struct inode *inode, int new, int size) inode->i_mode = type | (is_root_handle(inode) ? S_ISVTX : 0) | orangefs_inode_perms(&new_op->downcall.resp.getattr.attributes); + orangefs_inode->getattr_time = jiffies + getattr_timeout_msecs*HZ/1000; ret = 0; out: op_release(new_op); @@ -418,6 +421,7 @@ int orangefs_inode_setattr(struct inode *inode, struct iattr *iattr) ClearMtimeFlag(orangefs_inode); ClearCtimeFlag(orangefs_inode); ClearModeFlag(orangefs_inode); + orangefs_inode->getattr_time = jiffies - 1; } return ret; diff --git a/fs/orangefs/protocol.h b/fs/orangefs/protocol.h index 1efc6f8a5224..3d7418c728f5 100644 --- a/fs/orangefs/protocol.h +++ b/fs/orangefs/protocol.h @@ -207,14 +207,6 @@ typedef __s64 ORANGEFS_offset; ORANGEFS_ATTR_SYS_DIRENT_COUNT | \ ORANGEFS_ATTR_SYS_BLKSIZE) -#define ORANGEFS_ATTR_SYS_ALL_NOHINT_NOSIZE \ - (ORANGEFS_ATTR_SYS_COMMON_ALL | \ - ORANGEFS_ATTR_SYS_LNK_TARGET | \ - ORANGEFS_ATTR_SYS_DFILE_COUNT | \ - ORANGEFS_ATTR_SYS_MIRROR_COPIES_COUNT | \ - ORANGEFS_ATTR_SYS_DIRENT_COUNT | \ - ORANGEFS_ATTR_SYS_BLKSIZE) - #define ORANGEFS_XATTR_REPLACE 0x2 #define ORANGEFS_XATTR_CREATE 0x1 #define ORANGEFS_MAX_SERVER_ADDR_LEN 256 diff --git a/fs/proc/Makefile b/fs/proc/Makefile index 7151ea428041..a8c13605b434 100644 --- a/fs/proc/Makefile +++ b/fs/proc/Makefile @@ -4,6 +4,7 @@ obj-y += proc.o +CFLAGS_task_mmu.o += -Wno-override-init proc-y := nommu.o task_nommu.o proc-$(CONFIG_MMU) := task_mmu.o diff --git a/fs/proc/base.c b/fs/proc/base.c index 31370da2ee7c..54e270262979 100644 --- a/fs/proc/base.c +++ b/fs/proc/base.c @@ -579,11 +579,8 @@ static int proc_oom_score(struct seq_file *m, struct pid_namespace *ns, unsigned long totalpages = totalram_pages + total_swap_pages; unsigned long points = 0; - read_lock(&tasklist_lock); - if (pid_alive(task)) - points = oom_badness(task, NULL, NULL, totalpages) * - 1000 / totalpages; - read_unlock(&tasklist_lock); + points = oom_badness(task, NULL, NULL, totalpages) * + 1000 / totalpages; seq_printf(m, "%lu\n", points); return 0; diff --git a/fs/proc/stat.c b/fs/proc/stat.c index 510413eb25b8..7907e456ac4f 100644 --- a/fs/proc/stat.c +++ b/fs/proc/stat.c @@ -80,19 +80,17 @@ static u64 get_iowait_time(int cpu) static int show_stat(struct seq_file *p, void *v) { int i, j; - unsigned long jif; u64 user, nice, system, idle, iowait, irq, softirq, steal; u64 guest, guest_nice; u64 sum = 0; u64 sum_softirq = 0; unsigned int per_softirq_sums[NR_SOFTIRQS] = {0}; - struct timespec boottime; + struct timespec64 boottime; user = nice = system = idle = iowait = irq = softirq = steal = 0; guest = guest_nice = 0; - getboottime(&boottime); - jif = boottime.tv_sec; + getboottime64(&boottime); for_each_possible_cpu(i) { user += kcpustat_cpu(i).cpustat[CPUTIME_USER]; @@ -163,12 +161,12 @@ static int show_stat(struct seq_file *p, void *v) seq_printf(p, "\nctxt %llu\n" - "btime %lu\n" + "btime %llu\n" "processes %lu\n" "procs_running %lu\n" "procs_blocked %lu\n", nr_context_switches(), - (unsigned long)jif, + (unsigned long long)boottime.tv_sec, total_forks, nr_running(), nr_iowait()); diff --git a/fs/reiserfs/ibalance.c b/fs/reiserfs/ibalance.c index b751eea32e20..5db6f45b3fed 100644 --- a/fs/reiserfs/ibalance.c +++ b/fs/reiserfs/ibalance.c @@ -1153,8 +1153,9 @@ int balance_internal(struct tree_balance *tb, insert_ptr); } - memcpy(new_insert_key_addr, &new_insert_key, KEY_SIZE); insert_ptr[0] = new_insert_ptr; + if (new_insert_ptr) + memcpy(new_insert_key_addr, &new_insert_key, KEY_SIZE); return order; } |