diff options
Diffstat (limited to 'fs/ocfs2')
88 files changed, 5305 insertions, 3093 deletions
diff --git a/fs/ocfs2/Makefile b/fs/ocfs2/Makefile index d8a0313e99e..ce210d4951a 100644 --- a/fs/ocfs2/Makefile +++ b/fs/ocfs2/Makefile @@ -30,6 +30,7 @@ ocfs2-objs := \ namei.o \ refcounttree.o \ reservations.o \ + move_extents.o \ resize.o \ slot_map.o \ suballoc.o \ @@ -37,7 +38,6 @@ ocfs2-objs := \ symlink.o \ sysfile.o \ uptodate.o \ - ver.o \ quota_local.o \ quota_global.o \ xattr.o \ diff --git a/fs/ocfs2/acl.c b/fs/ocfs2/acl.c index e913ad130fd..7e8282dcea2 100644 --- a/fs/ocfs2/acl.c +++ b/fs/ocfs2/acl.c @@ -51,10 +51,6 @@ static struct posix_acl *ocfs2_acl_from_xattr(const void *value, size_t size) return ERR_PTR(-EINVAL); count = size / sizeof(struct posix_acl_entry); - if (count < 0) - return ERR_PTR(-EINVAL); - if (count == 0) - return NULL; acl = posix_acl_alloc(count, GFP_NOFS); if (!acl) @@ -65,7 +61,20 @@ static struct posix_acl *ocfs2_acl_from_xattr(const void *value, size_t size) acl->a_entries[n].e_tag = le16_to_cpu(entry->e_tag); acl->a_entries[n].e_perm = le16_to_cpu(entry->e_perm); - acl->a_entries[n].e_id = le32_to_cpu(entry->e_id); + switch(acl->a_entries[n].e_tag) { + case ACL_USER: + acl->a_entries[n].e_uid = + make_kuid(&init_user_ns, + le32_to_cpu(entry->e_id)); + break; + case ACL_GROUP: + acl->a_entries[n].e_gid = + make_kgid(&init_user_ns, + le32_to_cpu(entry->e_id)); + break; + default: + break; + } value += sizeof(struct posix_acl_entry); } @@ -91,7 +100,21 @@ static void *ocfs2_acl_to_xattr(const struct posix_acl *acl, size_t *size) for (n = 0; n < acl->a_count; n++, entry++) { entry->e_tag = cpu_to_le16(acl->a_entries[n].e_tag); entry->e_perm = cpu_to_le16(acl->a_entries[n].e_perm); - entry->e_id = cpu_to_le32(acl->a_entries[n].e_id); + switch(acl->a_entries[n].e_tag) { + case ACL_USER: + entry->e_id = cpu_to_le32( + from_kuid(&init_user_ns, + acl->a_entries[n].e_uid)); + break; + case ACL_GROUP: + entry->e_id = cpu_to_le32( + from_kgid(&init_user_ns, + acl->a_entries[n].e_gid)); + break; + default: + entry->e_id = cpu_to_le32(ACL_UNDEFINED_ID); + break; + } } return ocfs2_acl; } @@ -137,36 +160,6 @@ static struct posix_acl *ocfs2_get_acl_nolock(struct inode *inode, return acl; } - -/* - * Get posix acl. - */ -static struct posix_acl *ocfs2_get_acl(struct inode *inode, int type) -{ - struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); - struct buffer_head *di_bh = NULL; - struct posix_acl *acl; - int ret; - - if (!(osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL)) - return NULL; - - ret = ocfs2_inode_lock(inode, &di_bh, 0); - if (ret < 0) { - mlog_errno(ret); - acl = ERR_PTR(ret); - return acl; - } - - acl = ocfs2_get_acl_nolock(inode, type, di_bh); - - ocfs2_inode_unlock(inode, 0); - - brelse(di_bh); - - return acl; -} - /* * Helper function to set i_mode in memory and disk. Some call paths * will not have di_bh or a journal handle to pass, in which case it @@ -212,6 +205,7 @@ static int ocfs2_acl_set_mode(struct inode *inode, struct buffer_head *di_bh, di->i_mode = cpu_to_le16(inode->i_mode); di->i_ctime = cpu_to_le64(inode->i_ctime.tv_sec); di->i_ctime_nsec = cpu_to_le32(inode->i_ctime.tv_nsec); + ocfs2_update_inode_fsync_trans(handle, inode, 0); ocfs2_journal_dirty(handle, di_bh); @@ -227,7 +221,7 @@ out: /* * Set the access or default ACL of an inode. */ -static int ocfs2_set_acl(handle_t *handle, +int ocfs2_set_acl(handle_t *handle, struct inode *inode, struct buffer_head *di_bh, int type, @@ -247,7 +241,7 @@ static int ocfs2_set_acl(handle_t *handle, case ACL_TYPE_ACCESS: name_index = OCFS2_XATTR_INDEX_POSIX_ACL_ACCESS; if (acl) { - mode_t mode = inode->i_mode; + umode_t mode = inode->i_mode; ret = posix_acl_equiv_mode(acl, &mode); if (ret < 0) return ret; @@ -290,246 +284,29 @@ static int ocfs2_set_acl(handle_t *handle, return ret; } -int ocfs2_check_acl(struct inode *inode, int mask, unsigned int flags) +int ocfs2_iop_set_acl(struct inode *inode, struct posix_acl *acl, int type) +{ + return ocfs2_set_acl(NULL, inode, NULL, type, acl, NULL, NULL); +} + +struct posix_acl *ocfs2_iop_get_acl(struct inode *inode, int type) { struct ocfs2_super *osb; struct buffer_head *di_bh = NULL; struct posix_acl *acl; int ret = -EAGAIN; - if (flags & IPERM_FLAG_RCU) - return -ECHILD; - osb = OCFS2_SB(inode->i_sb); if (!(osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL)) - return ret; + return NULL; ret = ocfs2_read_inode_block(inode, &di_bh); - if (ret < 0) { - mlog_errno(ret); - return ret; - } + if (ret < 0) + return ERR_PTR(ret); - acl = ocfs2_get_acl_nolock(inode, ACL_TYPE_ACCESS, di_bh); + acl = ocfs2_get_acl_nolock(inode, type, di_bh); brelse(di_bh); - if (IS_ERR(acl)) { - mlog_errno(PTR_ERR(acl)); - return PTR_ERR(acl); - } - if (acl) { - ret = posix_acl_permission(inode, acl, mask); - posix_acl_release(acl); - return ret; - } - - return -EAGAIN; -} - -int ocfs2_acl_chmod(struct inode *inode) -{ - struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); - struct posix_acl *acl, *clone; - int ret; - - if (S_ISLNK(inode->i_mode)) - return -EOPNOTSUPP; - - if (!(osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL)) - return 0; - - acl = ocfs2_get_acl(inode, ACL_TYPE_ACCESS); - if (IS_ERR(acl) || !acl) - return PTR_ERR(acl); - clone = posix_acl_clone(acl, GFP_KERNEL); - posix_acl_release(acl); - if (!clone) - return -ENOMEM; - ret = posix_acl_chmod_masq(clone, inode->i_mode); - if (!ret) - ret = ocfs2_set_acl(NULL, inode, NULL, ACL_TYPE_ACCESS, - clone, NULL, NULL); - posix_acl_release(clone); - return ret; -} - -/* - * Initialize the ACLs of a new inode. If parent directory has default ACL, - * then clone to new inode. Called from ocfs2_mknod. - */ -int ocfs2_init_acl(handle_t *handle, - struct inode *inode, - struct inode *dir, - struct buffer_head *di_bh, - struct buffer_head *dir_bh, - struct ocfs2_alloc_context *meta_ac, - struct ocfs2_alloc_context *data_ac) -{ - struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); - struct posix_acl *acl = NULL; - int ret = 0, ret2; - mode_t mode; - - if (!S_ISLNK(inode->i_mode)) { - if (osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL) { - acl = ocfs2_get_acl_nolock(dir, ACL_TYPE_DEFAULT, - dir_bh); - if (IS_ERR(acl)) - return PTR_ERR(acl); - } - if (!acl) { - mode = inode->i_mode & ~current_umask(); - ret = ocfs2_acl_set_mode(inode, di_bh, handle, mode); - if (ret) { - mlog_errno(ret); - goto cleanup; - } - } - } - if ((osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL) && acl) { - struct posix_acl *clone; - - if (S_ISDIR(inode->i_mode)) { - ret = ocfs2_set_acl(handle, inode, di_bh, - ACL_TYPE_DEFAULT, acl, - meta_ac, data_ac); - if (ret) - goto cleanup; - } - clone = posix_acl_clone(acl, GFP_NOFS); - ret = -ENOMEM; - if (!clone) - goto cleanup; - - mode = inode->i_mode; - ret = posix_acl_create_masq(clone, &mode); - if (ret >= 0) { - ret2 = ocfs2_acl_set_mode(inode, di_bh, handle, mode); - if (ret2) { - mlog_errno(ret2); - ret = ret2; - goto cleanup; - } - if (ret > 0) { - ret = ocfs2_set_acl(handle, inode, - di_bh, ACL_TYPE_ACCESS, - clone, meta_ac, data_ac); - } - } - posix_acl_release(clone); - } -cleanup: - posix_acl_release(acl); - return ret; -} - -static size_t ocfs2_xattr_list_acl_access(struct dentry *dentry, - char *list, - size_t list_len, - const char *name, - size_t name_len, - int type) -{ - struct ocfs2_super *osb = OCFS2_SB(dentry->d_sb); - const size_t size = sizeof(POSIX_ACL_XATTR_ACCESS); - - if (!(osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL)) - return 0; - - if (list && size <= list_len) - memcpy(list, POSIX_ACL_XATTR_ACCESS, size); - return size; -} - -static size_t ocfs2_xattr_list_acl_default(struct dentry *dentry, - char *list, - size_t list_len, - const char *name, - size_t name_len, - int type) -{ - struct ocfs2_super *osb = OCFS2_SB(dentry->d_sb); - const size_t size = sizeof(POSIX_ACL_XATTR_DEFAULT); - - if (!(osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL)) - return 0; - - if (list && size <= list_len) - memcpy(list, POSIX_ACL_XATTR_DEFAULT, size); - return size; -} - -static int ocfs2_xattr_get_acl(struct dentry *dentry, const char *name, - void *buffer, size_t size, int type) -{ - struct ocfs2_super *osb = OCFS2_SB(dentry->d_sb); - struct posix_acl *acl; - int ret; - - if (strcmp(name, "") != 0) - return -EINVAL; - if (!(osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL)) - return -EOPNOTSUPP; - - acl = ocfs2_get_acl(dentry->d_inode, type); - if (IS_ERR(acl)) - return PTR_ERR(acl); - if (acl == NULL) - return -ENODATA; - ret = posix_acl_to_xattr(acl, buffer, size); - posix_acl_release(acl); - - return ret; -} - -static int ocfs2_xattr_set_acl(struct dentry *dentry, const char *name, - const void *value, size_t size, int flags, int type) -{ - struct inode *inode = dentry->d_inode; - struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); - struct posix_acl *acl; - int ret = 0; - - if (strcmp(name, "") != 0) - return -EINVAL; - if (!(osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL)) - return -EOPNOTSUPP; - - if (!inode_owner_or_capable(inode)) - return -EPERM; - - if (value) { - acl = posix_acl_from_xattr(value, size); - if (IS_ERR(acl)) - return PTR_ERR(acl); - else if (acl) { - ret = posix_acl_valid(acl); - if (ret) - goto cleanup; - } - } else - acl = NULL; - - ret = ocfs2_set_acl(NULL, inode, NULL, type, acl, NULL, NULL); - -cleanup: - posix_acl_release(acl); - return ret; + return acl; } - -const struct xattr_handler ocfs2_xattr_acl_access_handler = { - .prefix = POSIX_ACL_XATTR_ACCESS, - .flags = ACL_TYPE_ACCESS, - .list = ocfs2_xattr_list_acl_access, - .get = ocfs2_xattr_get_acl, - .set = ocfs2_xattr_set_acl, -}; - -const struct xattr_handler ocfs2_xattr_acl_default_handler = { - .prefix = POSIX_ACL_XATTR_DEFAULT, - .flags = ACL_TYPE_DEFAULT, - .list = ocfs2_xattr_list_acl_default, - .get = ocfs2_xattr_get_acl, - .set = ocfs2_xattr_set_acl, -}; diff --git a/fs/ocfs2/acl.h b/fs/ocfs2/acl.h index 4fe7c9cf4bf..3fce68d0862 100644 --- a/fs/ocfs2/acl.h +++ b/fs/ocfs2/acl.h @@ -26,11 +26,14 @@ struct ocfs2_acl_entry { __le32 e_id; }; -extern int ocfs2_check_acl(struct inode *, int, unsigned int); -extern int ocfs2_acl_chmod(struct inode *); -extern int ocfs2_init_acl(handle_t *, struct inode *, struct inode *, - struct buffer_head *, struct buffer_head *, - struct ocfs2_alloc_context *, - struct ocfs2_alloc_context *); +struct posix_acl *ocfs2_iop_get_acl(struct inode *inode, int type); +int ocfs2_iop_set_acl(struct inode *inode, struct posix_acl *acl, int type); +int ocfs2_set_acl(handle_t *handle, + struct inode *inode, + struct buffer_head *di_bh, + int type, + struct posix_acl *acl, + struct ocfs2_alloc_context *meta_ac, + struct ocfs2_alloc_context *data_ac); #endif /* OCFS2_ACL_H */ diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c index 48aa9c7401c..9d8fcf2f3b9 100644 --- a/fs/ocfs2/alloc.c +++ b/fs/ocfs2/alloc.c @@ -29,6 +29,7 @@ #include <linux/highmem.h> #include <linux/swap.h> #include <linux/quotaops.h> +#include <linux/blkdev.h> #include <cluster/masklog.h> @@ -1024,7 +1025,7 @@ static int ocfs2_create_new_meta_bhs(handle_t *handle, for(i = count; i < (num_got + count); i++) { bhs[i] = sb_getblk(osb->sb, first_blkno); if (bhs[i] == NULL) { - status = -EIO; + status = -ENOMEM; mlog_errno(status); goto bail; } @@ -1133,7 +1134,7 @@ static int ocfs2_adjust_rightmost_branch(handle_t *handle, } el = path_leaf_el(path); - rec = &el->l_recs[le32_to_cpu(el->l_next_free_rec) - 1]; + rec = &el->l_recs[le16_to_cpu(el->l_next_free_rec) - 1]; ocfs2_adjust_rightmost_records(handle, et, path, rec); @@ -4741,6 +4742,7 @@ int ocfs2_add_clusters_in_btree(handle_t *handle, enum ocfs2_alloc_restarted *reason_ret) { int status = 0, err = 0; + int need_free = 0; int free_extents; enum ocfs2_alloc_restarted reason = RESTART_NONE; u32 bit_off, num_bits; @@ -4795,7 +4797,8 @@ int ocfs2_add_clusters_in_btree(handle_t *handle, OCFS2_JOURNAL_ACCESS_WRITE); if (status < 0) { mlog_errno(status); - goto leave; + need_free = 1; + goto bail; } block = ocfs2_clusters_to_blocks(osb->sb, bit_off); @@ -4806,7 +4809,8 @@ int ocfs2_add_clusters_in_btree(handle_t *handle, num_bits, flags, meta_ac); if (status < 0) { mlog_errno(status); - goto leave; + need_free = 1; + goto bail; } ocfs2_journal_dirty(handle, et->et_root_bh); @@ -4820,6 +4824,19 @@ int ocfs2_add_clusters_in_btree(handle_t *handle, reason = RESTART_TRANS; } +bail: + if (need_free) { + if (data_ac->ac_which == OCFS2_AC_USE_LOCAL) + ocfs2_free_local_alloc_bits(osb, handle, data_ac, + bit_off, num_bits); + else + ocfs2_free_clusters(handle, + data_ac->ac_inode, + data_ac->ac_bh, + ocfs2_clusters_to_blocks(osb->sb, bit_off), + num_bits); + } + leave: if (reason_ret) *reason_ret = reason; @@ -5654,7 +5671,7 @@ int ocfs2_remove_btree_range(struct inode *inode, &ref_tree, NULL); if (ret) { mlog_errno(ret); - goto out; + goto bail; } ret = ocfs2_prepare_refcount_change_for_del(inode, @@ -5665,7 +5682,7 @@ int ocfs2_remove_btree_range(struct inode *inode, &extra_blocks); if (ret < 0) { mlog_errno(ret); - goto out; + goto bail; } } @@ -5673,7 +5690,7 @@ int ocfs2_remove_btree_range(struct inode *inode, extra_blocks); if (ret) { mlog_errno(ret); - return ret; + goto bail; } mutex_lock(&tl_inode->i_mutex); @@ -5698,7 +5715,7 @@ int ocfs2_remove_btree_range(struct inode *inode, OCFS2_JOURNAL_ACCESS_WRITE); if (ret) { mlog_errno(ret); - goto out; + goto out_commit; } dquot_free_space_nodirty(inode, @@ -5711,6 +5728,7 @@ int ocfs2_remove_btree_range(struct inode *inode, } ocfs2_et_update_clusters(et, -len); + ocfs2_update_inode_fsync_trans(handle, inode, 1); ocfs2_journal_dirty(handle, et->et_root_bh); @@ -5733,7 +5751,7 @@ out_commit: ocfs2_commit_trans(osb, handle); out: mutex_unlock(&tl_inode->i_mutex); - +bail: if (meta_ac) ocfs2_free_alloc_context(meta_ac); @@ -6028,7 +6046,8 @@ static void ocfs2_truncate_log_worker(struct work_struct *work) void ocfs2_schedule_truncate_log_flush(struct ocfs2_super *osb, int cancel) { - if (osb->osb_tl_inode) { + if (osb->osb_tl_inode && + atomic_read(&osb->osb_tl_disable) == 0) { /* We want to push off log flushes while truncates are * still running. */ if (cancel) @@ -6205,6 +6224,8 @@ void ocfs2_truncate_log_shutdown(struct ocfs2_super *osb) int status; struct inode *tl_inode = osb->osb_tl_inode; + atomic_set(&osb->osb_tl_disable, 1); + if (tl_inode) { cancel_delayed_work(&osb->osb_truncate_log_wq); flush_workqueue(ocfs2_wq); @@ -6236,6 +6257,7 @@ int ocfs2_truncate_log_init(struct ocfs2_super *osb) * until we're sure all is well. */ INIT_DELAYED_WORK(&osb->osb_truncate_log_wq, ocfs2_truncate_log_worker); + atomic_set(&osb->osb_tl_disable, 0); osb->osb_tl_bh = tl_bh; osb->osb_tl_inode = tl_inode; @@ -6750,8 +6772,7 @@ int ocfs2_zero_range_for_truncate(struct inode *inode, handle_t *handle, mlog_errno(ret); out: - if (pages) - kfree(pages); + kfree(pages); return ret; } @@ -6805,6 +6826,8 @@ int ocfs2_convert_inline_data_to_extents(struct inode *inode, struct buffer_head *di_bh) { int ret, i, has_data, num_pages = 0; + int need_free = 0; + u32 bit_off, num; handle_t *handle; u64 uninitialized_var(block); struct ocfs2_inode_info *oi = OCFS2_I(inode); @@ -6850,7 +6873,6 @@ int ocfs2_convert_inline_data_to_extents(struct inode *inode, } if (has_data) { - u32 bit_off, num; unsigned int page_end; u64 phys; @@ -6886,6 +6908,7 @@ int ocfs2_convert_inline_data_to_extents(struct inode *inode, ret = ocfs2_grab_eof_pages(inode, 0, end, pages, &num_pages); if (ret) { mlog_errno(ret); + need_free = 1; goto out_commit; } @@ -6896,6 +6919,7 @@ int ocfs2_convert_inline_data_to_extents(struct inode *inode, ret = ocfs2_read_inline_data(inode, pages[0], di_bh); if (ret) { mlog_errno(ret); + need_free = 1; goto out_commit; } @@ -6913,6 +6937,7 @@ int ocfs2_convert_inline_data_to_extents(struct inode *inode, di->i_dyn_features = cpu_to_le16(oi->ip_dyn_features); spin_unlock(&oi->ip_lock); + ocfs2_update_inode_fsync_trans(handle, inode, 1); ocfs2_dinode_new_extent_list(inode, di); ocfs2_journal_dirty(handle, di_bh); @@ -6927,6 +6952,7 @@ int ocfs2_convert_inline_data_to_extents(struct inode *inode, ret = ocfs2_insert_extent(handle, &et, 0, block, 1, 0, NULL); if (ret) { mlog_errno(ret); + need_free = 1; goto out_commit; } @@ -6938,6 +6964,18 @@ out_commit: dquot_free_space_nodirty(inode, ocfs2_clusters_to_bytes(osb->sb, 1)); + if (need_free) { + if (data_ac->ac_which == OCFS2_AC_USE_LOCAL) + ocfs2_free_local_alloc_bits(osb, handle, data_ac, + bit_off, num); + else + ocfs2_free_clusters(handle, + data_ac->ac_inode, + data_ac->ac_bh, + ocfs2_clusters_to_blocks(osb->sb, bit_off), + num); + } + ocfs2_commit_trans(osb, handle); out_unlock: @@ -7126,7 +7164,7 @@ int ocfs2_truncate_inline(struct inode *inode, struct buffer_head *di_bh, if (end > i_size_read(inode)) end = i_size_read(inode); - BUG_ON(start >= end); + BUG_ON(start > end); if (!(OCFS2_I(inode)->ip_dyn_features & OCFS2_INLINE_DATA_FL) || !(le16_to_cpu(di->i_dyn_features) & OCFS2_INLINE_DATA_FL) || @@ -7176,6 +7214,7 @@ int ocfs2_truncate_inline(struct inode *inode, struct buffer_head *di_bh, di->i_ctime = di->i_mtime = cpu_to_le64(inode->i_ctime.tv_sec); di->i_ctime_nsec = di->i_mtime_nsec = cpu_to_le32(inode->i_ctime.tv_nsec); + ocfs2_update_inode_fsync_trans(handle, inode, 1); ocfs2_journal_dirty(handle, di_bh); out_commit: @@ -7184,3 +7223,164 @@ out_commit: out: return ret; } + +static int ocfs2_trim_extent(struct super_block *sb, + struct ocfs2_group_desc *gd, + u32 start, u32 count) +{ + u64 discard, bcount; + + bcount = ocfs2_clusters_to_blocks(sb, count); + discard = le64_to_cpu(gd->bg_blkno) + + ocfs2_clusters_to_blocks(sb, start); + + trace_ocfs2_trim_extent(sb, (unsigned long long)discard, bcount); + + return sb_issue_discard(sb, discard, bcount, GFP_NOFS, 0); +} + +static int ocfs2_trim_group(struct super_block *sb, + struct ocfs2_group_desc *gd, + u32 start, u32 max, u32 minbits) +{ + int ret = 0, count = 0, next; + void *bitmap = gd->bg_bitmap; + + if (le16_to_cpu(gd->bg_free_bits_count) < minbits) + return 0; + + trace_ocfs2_trim_group((unsigned long long)le64_to_cpu(gd->bg_blkno), + start, max, minbits); + + while (start < max) { + start = ocfs2_find_next_zero_bit(bitmap, max, start); + if (start >= max) + break; + next = ocfs2_find_next_bit(bitmap, max, start); + + if ((next - start) >= minbits) { + ret = ocfs2_trim_extent(sb, gd, + start, next - start); + if (ret < 0) { + mlog_errno(ret); + break; + } + count += next - start; + } + start = next + 1; + + if (fatal_signal_pending(current)) { + count = -ERESTARTSYS; + break; + } + + if ((le16_to_cpu(gd->bg_free_bits_count) - count) < minbits) + break; + } + + if (ret < 0) + count = ret; + + return count; +} + +int ocfs2_trim_fs(struct super_block *sb, struct fstrim_range *range) +{ + struct ocfs2_super *osb = OCFS2_SB(sb); + u64 start, len, trimmed, first_group, last_group, group; + int ret, cnt; + u32 first_bit, last_bit, minlen; + struct buffer_head *main_bm_bh = NULL; + struct inode *main_bm_inode = NULL; + struct buffer_head *gd_bh = NULL; + struct ocfs2_dinode *main_bm; + struct ocfs2_group_desc *gd = NULL; + + start = range->start >> osb->s_clustersize_bits; + len = range->len >> osb->s_clustersize_bits; + minlen = range->minlen >> osb->s_clustersize_bits; + + if (minlen >= osb->bitmap_cpg || range->len < sb->s_blocksize) + return -EINVAL; + + main_bm_inode = ocfs2_get_system_file_inode(osb, + GLOBAL_BITMAP_SYSTEM_INODE, + OCFS2_INVALID_SLOT); + if (!main_bm_inode) { + ret = -EIO; + mlog_errno(ret); + goto out; + } + + mutex_lock(&main_bm_inode->i_mutex); + + ret = ocfs2_inode_lock(main_bm_inode, &main_bm_bh, 0); + if (ret < 0) { + mlog_errno(ret); + goto out_mutex; + } + main_bm = (struct ocfs2_dinode *)main_bm_bh->b_data; + + if (start >= le32_to_cpu(main_bm->i_clusters)) { + ret = -EINVAL; + goto out_unlock; + } + + len = range->len >> osb->s_clustersize_bits; + if (start + len > le32_to_cpu(main_bm->i_clusters)) + len = le32_to_cpu(main_bm->i_clusters) - start; + + trace_ocfs2_trim_fs(start, len, minlen); + + /* Determine first and last group to examine based on start and len */ + first_group = ocfs2_which_cluster_group(main_bm_inode, start); + if (first_group == osb->first_cluster_group_blkno) + first_bit = start; + else + first_bit = start - ocfs2_blocks_to_clusters(sb, first_group); + last_group = ocfs2_which_cluster_group(main_bm_inode, start + len - 1); + last_bit = osb->bitmap_cpg; + + trimmed = 0; + for (group = first_group; group <= last_group;) { + if (first_bit + len >= osb->bitmap_cpg) + last_bit = osb->bitmap_cpg; + else + last_bit = first_bit + len; + + ret = ocfs2_read_group_descriptor(main_bm_inode, + main_bm, group, + &gd_bh); + if (ret < 0) { + mlog_errno(ret); + break; + } + + gd = (struct ocfs2_group_desc *)gd_bh->b_data; + cnt = ocfs2_trim_group(sb, gd, first_bit, last_bit, minlen); + brelse(gd_bh); + gd_bh = NULL; + if (cnt < 0) { + ret = cnt; + mlog_errno(ret); + break; + } + + trimmed += cnt; + len -= osb->bitmap_cpg - first_bit; + first_bit = 0; + if (group == osb->first_cluster_group_blkno) + group = ocfs2_clusters_to_blocks(sb, osb->bitmap_cpg); + else + group += ocfs2_clusters_to_blocks(sb, osb->bitmap_cpg); + } + range->len = trimmed * sb->s_blocksize; +out_unlock: + ocfs2_inode_unlock(main_bm_inode, 0); + brelse(main_bm_bh); +out_mutex: + mutex_unlock(&main_bm_inode->i_mutex); + iput(main_bm_inode); +out: + return ret; +} diff --git a/fs/ocfs2/alloc.h b/fs/ocfs2/alloc.h index 3bd08a03251..ca381c58412 100644 --- a/fs/ocfs2/alloc.h +++ b/fs/ocfs2/alloc.h @@ -239,6 +239,7 @@ int ocfs2_find_leaf(struct ocfs2_caching_info *ci, struct buffer_head **leaf_bh); int ocfs2_search_extent_list(struct ocfs2_extent_list *el, u32 v_cluster); +int ocfs2_trim_fs(struct super_block *sb, struct fstrim_range *range); /* * Helper function to look at the # of clusters in an extent record. */ diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c index ac97bca282d..4a231a166cf 100644 --- a/fs/ocfs2/aops.c +++ b/fs/ocfs2/aops.c @@ -80,6 +80,7 @@ static int ocfs2_symlink_get_block(struct inode *inode, sector_t iblock, if ((u64)iblock >= ocfs2_clusters_to_blocks(inode->i_sb, le32_to_cpu(fe->i_clusters))) { + err = -ENOMEM; mlog(ML_ERROR, "block offset is outside the allocated size: " "%llu\n", (unsigned long long)iblock); goto bail; @@ -92,6 +93,7 @@ static int ocfs2_symlink_get_block(struct inode *inode, sector_t iblock, iblock; buffer_cache_bh = sb_getblk(osb->sb, blkno); if (!buffer_cache_bh) { + err = -ENOMEM; mlog(ML_ERROR, "couldn't getblock for symlink!\n"); goto bail; } @@ -102,7 +104,7 @@ static int ocfs2_symlink_get_block(struct inode *inode, sector_t iblock, * copy, the data is still good. */ if (buffer_jbd(buffer_cache_bh) && ocfs2_inode_is_new(inode)) { - kaddr = kmap_atomic(bh_result->b_page, KM_USER0); + kaddr = kmap_atomic(bh_result->b_page); if (!kaddr) { mlog(ML_ERROR, "couldn't kmap!\n"); goto bail; @@ -110,7 +112,7 @@ static int ocfs2_symlink_get_block(struct inode *inode, sector_t iblock, memcpy(kaddr + (bh_result->b_size * iblock), buffer_cache_bh->b_data, bh_result->b_size); - kunmap_atomic(kaddr, KM_USER0); + kunmap_atomic(kaddr); set_buffer_uptodate(bh_result); } brelse(buffer_cache_bh); @@ -236,13 +238,13 @@ int ocfs2_read_inline_data(struct inode *inode, struct page *page, return -EROFS; } - kaddr = kmap_atomic(page, KM_USER0); + kaddr = kmap_atomic(page); if (size) memcpy(kaddr, di->id2.i_data.id_data, size); /* Clear the remaining part of the page */ memset(kaddr + size, 0, PAGE_CACHE_SIZE - size); flush_dcache_page(page); - kunmap_atomic(kaddr, KM_USER0); + kunmap_atomic(kaddr); SetPageUptodate(page); @@ -290,7 +292,15 @@ static int ocfs2_readpage(struct file *file, struct page *page) } if (down_read_trylock(&oi->ip_alloc_sem) == 0) { + /* + * Unlock the page and cycle ip_alloc_sem so that we don't + * busyloop waiting for ip_alloc_sem to unlock + */ ret = AOP_TRUNCATED_PAGE; + unlock_page(page); + unlock = 0; + down_read(&oi->ip_alloc_sem); + up_read(&oi->ip_alloc_sem); goto out_inode_unlock; } @@ -551,66 +561,49 @@ bail: /* * ocfs2_dio_end_io is called by the dio core when a dio is finished. We're - * particularly interested in the aio/dio case. Like the core uses - * i_alloc_sem, we use the rw_lock DLM lock to protect io on one node from - * truncation on another. + * particularly interested in the aio/dio case. We use the rw_lock DLM lock + * to protect io on one node from truncation on another. */ static void ocfs2_dio_end_io(struct kiocb *iocb, loff_t offset, ssize_t bytes, - void *private, - int ret, - bool is_async) + void *private) { - struct inode *inode = iocb->ki_filp->f_path.dentry->d_inode; + struct inode *inode = file_inode(iocb->ki_filp); int level; /* this io's submitter should not have unlocked this before we could */ BUG_ON(!ocfs2_iocb_is_rw_locked(iocb)); - if (ocfs2_iocb_is_sem_locked(iocb)) { - up_read(&inode->i_alloc_sem); + if (ocfs2_iocb_is_sem_locked(iocb)) ocfs2_iocb_clear_sem_locked(iocb); + + if (ocfs2_iocb_is_unaligned_aio(iocb)) { + ocfs2_iocb_clear_unaligned_aio(iocb); + + mutex_unlock(&OCFS2_I(inode)->ip_unaligned_aio); } ocfs2_iocb_clear_rw_locked(iocb); level = ocfs2_iocb_rw_locked_level(iocb); ocfs2_rw_unlock(inode, level); - - if (is_async) - aio_complete(iocb, ret, 0); -} - -/* - * ocfs2_invalidatepage() and ocfs2_releasepage() are shamelessly stolen - * from ext3. PageChecked() bits have been removed as OCFS2 does not - * do journalled data. - */ -static void ocfs2_invalidatepage(struct page *page, unsigned long offset) -{ - journal_t *journal = OCFS2_SB(page->mapping->host->i_sb)->journal->j_journal; - - jbd2_journal_invalidatepage(journal, page, offset); } static int ocfs2_releasepage(struct page *page, gfp_t wait) { - journal_t *journal = OCFS2_SB(page->mapping->host->i_sb)->journal->j_journal; - if (!page_has_buffers(page)) return 0; - return jbd2_journal_try_to_free_buffers(journal, page, wait); + return try_to_free_buffers(page); } static ssize_t ocfs2_direct_IO(int rw, struct kiocb *iocb, - const struct iovec *iov, - loff_t offset, - unsigned long nr_segs) + struct iov_iter *iter, + loff_t offset) { struct file *file = iocb->ki_filp; - struct inode *inode = file->f_path.dentry->d_inode->i_mapping->host; + struct inode *inode = file_inode(file)->i_mapping->host; /* * Fallback to buffered I/O if we see an inode without @@ -624,7 +617,7 @@ static ssize_t ocfs2_direct_IO(int rw, return 0; return __blockdev_direct_IO(rw, iocb, inode, inode->i_sb->s_bdev, - iov, offset, nr_segs, + iter, offset, ocfs2_direct_IO_get_blocks, ocfs2_dio_end_io, NULL, 0); } @@ -673,7 +666,7 @@ static void ocfs2_clear_page_regions(struct page *page, ocfs2_figure_cluster_boundaries(osb, cpos, &cluster_start, &cluster_end); - kaddr = kmap_atomic(page, KM_USER0); + kaddr = kmap_atomic(page); if (from || to) { if (from > cluster_start) @@ -684,7 +677,7 @@ static void ocfs2_clear_page_regions(struct page *page, memset(kaddr + cluster_start, 0, cluster_end - cluster_start); } - kunmap_atomic(kaddr, KM_USER0); + kunmap_atomic(kaddr); } /* @@ -865,6 +858,12 @@ struct ocfs2_write_ctxt { struct page *w_target_page; /* + * w_target_locked is used for page_mkwrite path indicating no unlocking + * against w_target_page in ocfs2_write_end_nolock. + */ + unsigned int w_target_locked:1; + + /* * ocfs2_write_end() uses this to know what the real range to * write in the target should be. */ @@ -897,6 +896,24 @@ void ocfs2_unlock_and_free_pages(struct page **pages, int num_pages) static void ocfs2_free_write_ctxt(struct ocfs2_write_ctxt *wc) { + int i; + + /* + * w_target_locked is only set to true in the page_mkwrite() case. + * The intent is to allow us to lock the target page from write_begin() + * to write_end(). The caller must hold a ref on w_target_page. + */ + if (wc->w_target_locked) { + BUG_ON(!wc->w_target_page); + for (i = 0; i < wc->w_num_pages; i++) { + if (wc->w_target_page == wc->w_pages[i]) { + wc->w_pages[i] = NULL; + break; + } + } + mark_page_accessed(wc->w_target_page); + page_cache_release(wc->w_target_page); + } ocfs2_unlock_and_free_pages(wc->w_pages, wc->w_num_pages); brelse(wc->w_di_bh); @@ -1134,20 +1151,17 @@ static int ocfs2_grab_pages_for_write(struct address_space *mapping, */ lock_page(mmap_page); + /* Exit and let the caller retry */ if (mmap_page->mapping != mapping) { + WARN_ON(mmap_page->mapping); unlock_page(mmap_page); - /* - * Sanity check - the locking in - * ocfs2_pagemkwrite() should ensure - * that this code doesn't trigger. - */ - ret = -EINVAL; - mlog_errno(ret); + ret = -EAGAIN; goto out; } page_cache_get(mmap_page); wc->w_pages[i] = mmap_page; + wc->w_target_locked = true; } else { wc->w_pages[i] = find_or_create_page(mapping, index, GFP_NOFS); @@ -1157,11 +1171,14 @@ static int ocfs2_grab_pages_for_write(struct address_space *mapping, goto out; } } + wait_for_stable_page(wc->w_pages[i]); if (index == target_index) wc->w_target_page = wc->w_pages[i]; } out: + if (ret) + wc->w_target_locked = false; return ret; } @@ -1716,7 +1733,7 @@ try_again: goto out; } else if (ret == 1) { clusters_need = wc->w_clen; - ret = ocfs2_refcount_cow(inode, filp, di_bh, + ret = ocfs2_refcount_cow(inode, di_bh, wc->w_cpos, wc->w_clen, UINT_MAX); if (ret) { mlog_errno(ret); @@ -1767,8 +1784,7 @@ try_again: data_ac->ac_resv = &OCFS2_I(inode)->ip_la_data_resv; credits = ocfs2_calc_extend_credits(inode->i_sb, - &di->id2.i_list, - clusters_to_alloc); + &di->id2.i_list); } @@ -1819,11 +1835,23 @@ try_again: */ ret = ocfs2_grab_pages_for_write(mapping, wc, wc->w_cpos, pos, len, cluster_of_pages, mmap_page); - if (ret) { + if (ret && ret != -EAGAIN) { mlog_errno(ret); goto out_quota; } + /* + * ocfs2_grab_pages_for_write() returns -EAGAIN if it could not lock + * the target page. In this case, we exit with no error and no target + * page. This will trigger the caller, page_mkwrite(), to re-try + * the operation. + */ + if (ret == -EAGAIN) { + BUG_ON(wc->w_target_page); + ret = 0; + goto out_quota; + } + ret = ocfs2_write_cluster_by_desc(mapping, data_ac, meta_ac, wc, pos, len); if (ret) { @@ -1850,10 +1878,14 @@ out_commit: out: ocfs2_free_write_ctxt(wc); - if (data_ac) + if (data_ac) { ocfs2_free_alloc_context(data_ac); - if (meta_ac) + data_ac = NULL; + } + if (meta_ac) { ocfs2_free_alloc_context(meta_ac); + meta_ac = NULL; + } if (ret == -ENOSPC && try_free) { /* @@ -1930,9 +1962,9 @@ static void ocfs2_write_end_inline(struct inode *inode, loff_t pos, } } - kaddr = kmap_atomic(wc->w_target_page, KM_USER0); + kaddr = kmap_atomic(wc->w_target_page); memcpy(di->id2.i_data.id_data + pos, kaddr + pos, *copied); - kunmap_atomic(kaddr, KM_USER0); + kunmap_atomic(kaddr); trace_ocfs2_write_end_inline( (unsigned long long)OCFS2_I(inode)->ip_blkno, @@ -1997,7 +2029,7 @@ int ocfs2_write_end_nolock(struct address_space *mapping, out_write_size: pos += copied; - if (pos > inode->i_size) { + if (pos > i_size_read(inode)) { i_size_write(inode, pos); mark_inode_dirty(inode); } @@ -2006,6 +2038,7 @@ out_write_size: inode->i_mtime = inode->i_ctime = CURRENT_TIME; di->i_mtime = di->i_ctime = cpu_to_le64(inode->i_mtime.tv_sec); di->i_mtime_nsec = di->i_ctime_nsec = cpu_to_le32(inode->i_mtime.tv_nsec); + ocfs2_update_inode_fsync_trans(handle, inode, 1); ocfs2_journal_dirty(handle, wc->w_di_bh); ocfs2_commit_trans(osb, handle); @@ -2040,7 +2073,7 @@ const struct address_space_operations ocfs2_aops = { .write_end = ocfs2_write_end, .bmap = ocfs2_bmap, .direct_IO = ocfs2_direct_IO, - .invalidatepage = ocfs2_invalidatepage, + .invalidatepage = block_invalidatepage, .releasepage = ocfs2_releasepage, .migratepage = buffer_migrate_page, .is_partially_uptodate = block_is_partially_uptodate, diff --git a/fs/ocfs2/aops.h b/fs/ocfs2/aops.h index 75cf3ad987a..6cae155d54d 100644 --- a/fs/ocfs2/aops.h +++ b/fs/ocfs2/aops.h @@ -22,6 +22,8 @@ #ifndef OCFS2_AOPS_H #define OCFS2_AOPS_H +#include <linux/aio.h> + handle_t *ocfs2_start_walk_page_trans(struct inode *inode, struct page *page, unsigned from, @@ -78,6 +80,7 @@ enum ocfs2_iocb_lock_bits { OCFS2_IOCB_RW_LOCK = 0, OCFS2_IOCB_RW_LOCK_LEVEL, OCFS2_IOCB_SEM, + OCFS2_IOCB_UNALIGNED_IO, OCFS2_IOCB_NUM_LOCKS }; @@ -91,4 +94,12 @@ enum ocfs2_iocb_lock_bits { clear_bit(OCFS2_IOCB_SEM, (unsigned long *)&iocb->private) #define ocfs2_iocb_is_sem_locked(iocb) \ test_bit(OCFS2_IOCB_SEM, (unsigned long *)&iocb->private) + +#define ocfs2_iocb_set_unaligned_aio(iocb) \ + set_bit(OCFS2_IOCB_UNALIGNED_IO, (unsigned long *)&iocb->private) +#define ocfs2_iocb_clear_unaligned_aio(iocb) \ + clear_bit(OCFS2_IOCB_UNALIGNED_IO, (unsigned long *)&iocb->private) +#define ocfs2_iocb_is_unaligned_aio(iocb) \ + test_bit(OCFS2_IOCB_UNALIGNED_IO, (unsigned long *)&iocb->private) + #endif /* OCFS2_FILE_H */ diff --git a/fs/ocfs2/blockcheck.c b/fs/ocfs2/blockcheck.c index c7ee03c2222..0725e605465 100644 --- a/fs/ocfs2/blockcheck.c +++ b/fs/ocfs2/blockcheck.c @@ -422,45 +422,46 @@ int ocfs2_block_check_validate(void *data, size_t blocksize, struct ocfs2_blockcheck_stats *stats) { int rc = 0; - struct ocfs2_block_check check; + u32 bc_crc32e; + u16 bc_ecc; u32 crc, ecc; ocfs2_blockcheck_inc_check(stats); - check.bc_crc32e = le32_to_cpu(bc->bc_crc32e); - check.bc_ecc = le16_to_cpu(bc->bc_ecc); + bc_crc32e = le32_to_cpu(bc->bc_crc32e); + bc_ecc = le16_to_cpu(bc->bc_ecc); memset(bc, 0, sizeof(struct ocfs2_block_check)); /* Fast path - if the crc32 validates, we're good to go */ crc = crc32_le(~0, data, blocksize); - if (crc == check.bc_crc32e) + if (crc == bc_crc32e) goto out; ocfs2_blockcheck_inc_failure(stats); mlog(ML_ERROR, "CRC32 failed: stored: 0x%x, computed 0x%x. Applying ECC.\n", - (unsigned int)check.bc_crc32e, (unsigned int)crc); + (unsigned int)bc_crc32e, (unsigned int)crc); /* Ok, try ECC fixups */ ecc = ocfs2_hamming_encode_block(data, blocksize); - ocfs2_hamming_fix_block(data, blocksize, ecc ^ check.bc_ecc); + ocfs2_hamming_fix_block(data, blocksize, ecc ^ bc_ecc); /* And check the crc32 again */ crc = crc32_le(~0, data, blocksize); - if (crc == check.bc_crc32e) { + if (crc == bc_crc32e) { ocfs2_blockcheck_inc_recover(stats); goto out; } mlog(ML_ERROR, "Fixed CRC32 failed: stored: 0x%x, computed 0x%x\n", - (unsigned int)check.bc_crc32e, (unsigned int)crc); + (unsigned int)bc_crc32e, (unsigned int)crc); rc = -EIO; out: - bc->bc_crc32e = cpu_to_le32(check.bc_crc32e); - bc->bc_ecc = cpu_to_le16(check.bc_ecc); + bc->bc_crc32e = cpu_to_le32(bc_crc32e); + bc->bc_ecc = cpu_to_le16(bc_ecc); return rc; } @@ -528,7 +529,8 @@ int ocfs2_block_check_validate_bhs(struct buffer_head **bhs, int nr, struct ocfs2_blockcheck_stats *stats) { int i, rc = 0; - struct ocfs2_block_check check; + u32 bc_crc32e; + u16 bc_ecc; u32 crc, ecc, fix; BUG_ON(nr < 0); @@ -538,21 +540,21 @@ int ocfs2_block_check_validate_bhs(struct buffer_head **bhs, int nr, ocfs2_blockcheck_inc_check(stats); - check.bc_crc32e = le32_to_cpu(bc->bc_crc32e); - check.bc_ecc = le16_to_cpu(bc->bc_ecc); + bc_crc32e = le32_to_cpu(bc->bc_crc32e); + bc_ecc = le16_to_cpu(bc->bc_ecc); memset(bc, 0, sizeof(struct ocfs2_block_check)); /* Fast path - if the crc32 validates, we're good to go */ for (i = 0, crc = ~0; i < nr; i++) crc = crc32_le(crc, bhs[i]->b_data, bhs[i]->b_size); - if (crc == check.bc_crc32e) + if (crc == bc_crc32e) goto out; ocfs2_blockcheck_inc_failure(stats); mlog(ML_ERROR, "CRC32 failed: stored: %u, computed %u. Applying ECC.\n", - (unsigned int)check.bc_crc32e, (unsigned int)crc); + (unsigned int)bc_crc32e, (unsigned int)crc); /* Ok, try ECC fixups */ for (i = 0, ecc = 0; i < nr; i++) { @@ -565,7 +567,7 @@ int ocfs2_block_check_validate_bhs(struct buffer_head **bhs, int nr, bhs[i]->b_size * 8, bhs[i]->b_size * 8 * i); } - fix = ecc ^ check.bc_ecc; + fix = ecc ^ bc_ecc; for (i = 0; i < nr; i++) { /* * Try the fix against each buffer. It will only affect @@ -578,19 +580,19 @@ int ocfs2_block_check_validate_bhs(struct buffer_head **bhs, int nr, /* And check the crc32 again */ for (i = 0, crc = ~0; i < nr; i++) crc = crc32_le(crc, bhs[i]->b_data, bhs[i]->b_size); - if (crc == check.bc_crc32e) { + if (crc == bc_crc32e) { ocfs2_blockcheck_inc_recover(stats); goto out; } mlog(ML_ERROR, "Fixed CRC32 failed: stored: %u, computed %u\n", - (unsigned int)check.bc_crc32e, (unsigned int)crc); + (unsigned int)bc_crc32e, (unsigned int)crc); rc = -EIO; out: - bc->bc_crc32e = cpu_to_le32(check.bc_crc32e); - bc->bc_ecc = cpu_to_le16(check.bc_ecc); + bc->bc_crc32e = cpu_to_le32(bc_crc32e); + bc->bc_ecc = cpu_to_le16(bc_ecc); return rc; } diff --git a/fs/ocfs2/buffer_head_io.c b/fs/ocfs2/buffer_head_io.c index 5d18ad10c27..1edcb141f63 100644 --- a/fs/ocfs2/buffer_head_io.c +++ b/fs/ocfs2/buffer_head_io.c @@ -90,7 +90,6 @@ int ocfs2_write_block(struct ocfs2_super *osb, struct buffer_head *bh, * information for this bh as it's not marked locally * uptodate. */ ret = -EIO; - put_bh(bh); mlog_errno(ret); } @@ -115,7 +114,7 @@ int ocfs2_read_blocks_sync(struct ocfs2_super *osb, u64 block, if (bhs[i] == NULL) { bhs[i] = sb_getblk(osb->sb, block++); if (bhs[i] == NULL) { - status = -EIO; + status = -ENOMEM; mlog_errno(status); goto bail; } @@ -214,7 +213,7 @@ int ocfs2_read_blocks(struct ocfs2_caching_info *ci, u64 block, int nr, bhs[i] = sb_getblk(sb, block++); if (bhs[i] == NULL) { ocfs2_metadata_cache_io_unlock(ci); - status = -EIO; + status = -ENOMEM; mlog_errno(status); goto bail; } @@ -420,7 +419,6 @@ int ocfs2_write_super_or_backup(struct ocfs2_super *osb, if (!buffer_uptodate(bh)) { ret = -EIO; - put_bh(bh); mlog_errno(ret); } diff --git a/fs/ocfs2/cluster/Makefile b/fs/ocfs2/cluster/Makefile index bc8c5e7d860..1aefc0350ec 100644 --- a/fs/ocfs2/cluster/Makefile +++ b/fs/ocfs2/cluster/Makefile @@ -1,4 +1,4 @@ obj-$(CONFIG_OCFS2_FS) += ocfs2_nodemanager.o ocfs2_nodemanager-objs := heartbeat.o masklog.o sys.o nodemanager.o \ - quorum.o tcp.o netdebug.o ver.o + quorum.o tcp.o netdebug.o diff --git a/fs/ocfs2/cluster/heartbeat.c b/fs/ocfs2/cluster/heartbeat.c index 643720209a9..73039295d0d 100644 --- a/fs/ocfs2/cluster/heartbeat.c +++ b/fs/ocfs2/cluster/heartbeat.c @@ -35,6 +35,7 @@ #include <linux/time.h> #include <linux/debugfs.h> #include <linux/slab.h> +#include <linux/bitmap.h> #include "heartbeat.h" #include "tcp.h" @@ -176,7 +177,7 @@ static void o2hb_dead_threshold_set(unsigned int threshold) } } -static int o2hb_global_hearbeat_mode_set(unsigned int hb_mode) +static int o2hb_global_heartbeat_mode_set(unsigned int hb_mode) { int ret = -1; @@ -216,6 +217,7 @@ struct o2hb_region { struct list_head hr_all_item; unsigned hr_unclean_stop:1, + hr_aborted_start:1, hr_item_pinned:1, hr_item_dropped:1; @@ -254,6 +256,10 @@ struct o2hb_region { * a more complete api that doesn't lead to this sort of fragility. */ atomic_t hr_steady_iterations; + /* terminate o2hb thread if it does not reach steady state + * (hr_steady_iterations == 0) within hr_unsteady_iterations */ + atomic_t hr_unsteady_iterations; + char hr_dev_name[BDEVNAME_SIZE]; unsigned int hr_timeout_ms; @@ -277,15 +283,6 @@ struct o2hb_bio_wait_ctxt { int wc_error; }; -static int o2hb_pop_count(void *map, int count) -{ - int i = -1, pop = 0; - - while ((i = find_next_bit(map, count, i + 1)) < count) - pop++; - return pop; -} - static void o2hb_write_timeout(struct work_struct *work) { int failed, quorum; @@ -302,9 +299,9 @@ static void o2hb_write_timeout(struct work_struct *work) spin_lock_irqsave(&o2hb_live_lock, flags); if (test_bit(reg->hr_region_num, o2hb_quorum_region_bitmap)) set_bit(reg->hr_region_num, o2hb_failed_region_bitmap); - failed = o2hb_pop_count(&o2hb_failed_region_bitmap, + failed = bitmap_weight(o2hb_failed_region_bitmap, O2NM_MAX_REGIONS); - quorum = o2hb_pop_count(&o2hb_quorum_region_bitmap, + quorum = bitmap_weight(o2hb_quorum_region_bitmap, O2NM_MAX_REGIONS); spin_unlock_irqrestore(&o2hb_live_lock, flags); @@ -324,6 +321,10 @@ static void o2hb_write_timeout(struct work_struct *work) static void o2hb_arm_write_timeout(struct o2hb_region *reg) { + /* Arm writeout only after thread reaches steady state */ + if (atomic_read(®->hr_steady_iterations) != 0) + return; + mlog(ML_HEARTBEAT, "Queue write timeout for %u ms\n", O2HB_MAX_WRITE_TIMEOUT_MS); @@ -412,7 +413,7 @@ static struct bio *o2hb_setup_one_bio(struct o2hb_region *reg, } /* Must put everything in 512 byte sectors for the bio... */ - bio->bi_sector = (reg->hr_start_block + cs) << (bits - 9); + bio->bi_iter.bi_sector = (reg->hr_start_block + cs) << (bits - 9); bio->bi_bdev = reg->hr_bdev; bio->bi_private = wc; bio->bi_end_io = o2hb_bio_end_io; @@ -491,7 +492,7 @@ static int o2hb_issue_node_write(struct o2hb_region *reg, } atomic_inc(&write_wc->wc_num_reqs); - submit_bio(WRITE, bio); + submit_bio(WRITE_SYNC, bio); status = 0; bail: @@ -537,27 +538,50 @@ static int o2hb_verify_crc(struct o2hb_region *reg, return read == computed; } -/* We want to make sure that nobody is heartbeating on top of us -- - * this will help detect an invalid configuration. */ -static int o2hb_check_last_timestamp(struct o2hb_region *reg) +/* + * Compare the slot data with what we wrote in the last iteration. + * If the match fails, print an appropriate error message. This is to + * detect errors like... another node hearting on the same slot, + * flaky device that is losing writes, etc. + * Returns 1 if check succeeds, 0 otherwise. + */ +static int o2hb_check_own_slot(struct o2hb_region *reg) { - int node_num, ret; struct o2hb_disk_slot *slot; struct o2hb_disk_heartbeat_block *hb_block; + char *errstr; - node_num = o2nm_this_node(); - - ret = 1; - slot = ®->hr_slots[node_num]; + slot = ®->hr_slots[o2nm_this_node()]; /* Don't check on our 1st timestamp */ - if (slot->ds_last_time) { - hb_block = slot->ds_raw_block; + if (!slot->ds_last_time) + return 0; - if (le64_to_cpu(hb_block->hb_seq) != slot->ds_last_time) - ret = 0; - } + hb_block = slot->ds_raw_block; + if (le64_to_cpu(hb_block->hb_seq) == slot->ds_last_time && + le64_to_cpu(hb_block->hb_generation) == slot->ds_last_generation && + hb_block->hb_node == slot->ds_node_num) + return 1; + +#define ERRSTR1 "Another node is heartbeating on device" +#define ERRSTR2 "Heartbeat generation mismatch on device" +#define ERRSTR3 "Heartbeat sequence mismatch on device" + + if (hb_block->hb_node != slot->ds_node_num) + errstr = ERRSTR1; + else if (le64_to_cpu(hb_block->hb_generation) != + slot->ds_last_generation) + errstr = ERRSTR2; + else + errstr = ERRSTR3; - return ret; + mlog(ML_ERROR, "%s (%s): expected(%u:0x%llx, 0x%llx), " + "ondisk(%u:0x%llx, 0x%llx)\n", errstr, reg->hr_dev_name, + slot->ds_node_num, (unsigned long long)slot->ds_last_generation, + (unsigned long long)slot->ds_last_time, hb_block->hb_node, + (unsigned long long)le64_to_cpu(hb_block->hb_generation), + (unsigned long long)le64_to_cpu(hb_block->hb_seq)); + + return 0; } static inline void o2hb_prepare_block(struct o2hb_region *reg, @@ -596,11 +620,9 @@ static void o2hb_fire_callbacks(struct o2hb_callback *hbcall, struct o2nm_node *node, int idx) { - struct list_head *iter; struct o2hb_callback_func *f; - list_for_each(iter, &hbcall->list) { - f = list_entry(iter, struct o2hb_callback_func, hc_item); + list_for_each_entry(f, &hbcall->list, hc_item) { mlog(ML_HEARTBEAT, "calling funcs %p\n", f); (f->hc_func)(node, idx, f->hc_data); } @@ -609,16 +631,9 @@ static void o2hb_fire_callbacks(struct o2hb_callback *hbcall, /* Will run the list in order until we process the passed event */ static void o2hb_run_event_list(struct o2hb_node_event *queued_event) { - int empty; struct o2hb_callback *hbcall; struct o2hb_node_event *event; - spin_lock(&o2hb_live_lock); - empty = list_empty(&queued_event->hn_item); - spin_unlock(&o2hb_live_lock); - if (empty) - return; - /* Holding callback sem assures we don't alter the callback * lists when doing this, and serializes ourselves with other * processes wanting callbacks. */ @@ -677,6 +692,7 @@ static void o2hb_shutdown_slot(struct o2hb_disk_slot *slot) struct o2hb_node_event event = { .hn_item = LIST_HEAD_INIT(event.hn_item), }; struct o2nm_node *node; + int queued = 0; node = o2nm_get_node_by_num(slot->ds_node_num); if (!node) @@ -694,26 +710,35 @@ static void o2hb_shutdown_slot(struct o2hb_disk_slot *slot) o2hb_queue_node_event(&event, O2HB_NODE_DOWN_CB, node, slot->ds_node_num); + queued = 1; } } spin_unlock(&o2hb_live_lock); - o2hb_run_event_list(&event); + if (queued) + o2hb_run_event_list(&event); o2nm_node_put(node); } -static void o2hb_set_quorum_device(struct o2hb_region *reg, - struct o2hb_disk_slot *slot) +static void o2hb_set_quorum_device(struct o2hb_region *reg) { - assert_spin_locked(&o2hb_live_lock); - if (!o2hb_global_heartbeat_active()) return; - if (test_bit(reg->hr_region_num, o2hb_quorum_region_bitmap)) + /* Prevent race with o2hb_heartbeat_group_drop_item() */ + if (kthread_should_stop()) return; + /* Tag region as quorum only after thread reaches steady state */ + if (atomic_read(®->hr_steady_iterations) != 0) + return; + + spin_lock(&o2hb_live_lock); + + if (test_bit(reg->hr_region_num, o2hb_quorum_region_bitmap)) + goto unlock; + /* * A region can be added to the quorum only when it sees all * live nodes heartbeat on it. In other words, the region has been @@ -721,13 +746,10 @@ static void o2hb_set_quorum_device(struct o2hb_region *reg, */ if (memcmp(reg->hr_live_node_bitmap, o2hb_live_node_bitmap, sizeof(o2hb_live_node_bitmap))) - return; - - if (slot->ds_changed_samples < O2HB_LIVE_THRESHOLD) - return; + goto unlock; - printk(KERN_NOTICE "o2hb: Region %s is now a quorum device\n", - config_item_name(®->hr_item)); + printk(KERN_NOTICE "o2hb: Region %s (%s) is now a quorum device\n", + config_item_name(®->hr_item), reg->hr_dev_name); set_bit(reg->hr_region_num, o2hb_quorum_region_bitmap); @@ -735,9 +757,11 @@ static void o2hb_set_quorum_device(struct o2hb_region *reg, * If global heartbeat active, unpin all regions if the * region count > CUT_OFF */ - if (o2hb_pop_count(&o2hb_quorum_region_bitmap, + if (bitmap_weight(o2hb_quorum_region_bitmap, O2NM_MAX_REGIONS) > O2HB_PIN_CUT_OFF) o2hb_region_unpin(NULL); +unlock: + spin_unlock(&o2hb_live_lock); } static int o2hb_check_slot(struct o2hb_region *reg, @@ -752,6 +776,7 @@ static int o2hb_check_slot(struct o2hb_region *reg, unsigned int dead_ms = o2hb_dead_threshold * O2HB_REGION_TIMEOUT_MS; unsigned int slot_dead_ms; int tmp; + int queued = 0; memcpy(hb_block, slot->ds_raw_block, reg->hr_block_bytes); @@ -845,6 +870,7 @@ fire_callbacks: slot->ds_node_num); changed = 1; + queued = 1; } list_add_tail(&slot->ds_live_item, @@ -896,6 +922,7 @@ fire_callbacks: node, slot->ds_node_num); changed = 1; + queued = 1; } /* We don't clear this because the node is still @@ -909,39 +936,25 @@ fire_callbacks: slot->ds_equal_samples = 0; } out: - o2hb_set_quorum_device(reg, slot); - spin_unlock(&o2hb_live_lock); - o2hb_run_event_list(&event); + if (queued) + o2hb_run_event_list(&event); if (node) o2nm_node_put(node); return changed; } -/* This could be faster if we just implmented a find_last_bit, but I - * don't think the circumstances warrant it. */ -static int o2hb_highest_node(unsigned long *nodes, - int numbits) +static int o2hb_highest_node(unsigned long *nodes, int numbits) { - int highest, node; - - highest = numbits; - node = -1; - while ((node = find_next_bit(nodes, numbits, node + 1)) != -1) { - if (node >= numbits) - break; - - highest = node; - } - - return highest; + return find_last_bit(nodes, numbits); } static int o2hb_do_disk_heartbeat(struct o2hb_region *reg) { - int i, ret, highest_node, change = 0; + int i, ret, highest_node; + int membership_change = 0, own_slot_ok = 0; unsigned long configured_nodes[BITS_TO_LONGS(O2NM_MAX_NODES)]; unsigned long live_node_bitmap[BITS_TO_LONGS(O2NM_MAX_NODES)]; struct o2hb_bio_wait_ctxt write_wc; @@ -950,7 +963,7 @@ static int o2hb_do_disk_heartbeat(struct o2hb_region *reg) sizeof(configured_nodes)); if (ret) { mlog_errno(ret); - return ret; + goto bail; } /* @@ -966,8 +979,9 @@ static int o2hb_do_disk_heartbeat(struct o2hb_region *reg) highest_node = o2hb_highest_node(configured_nodes, O2NM_MAX_NODES); if (highest_node >= O2NM_MAX_NODES) { - mlog(ML_NOTICE, "ocfs2_heartbeat: no configured nodes found!\n"); - return -EINVAL; + mlog(ML_NOTICE, "o2hb: No configured nodes found!\n"); + ret = -EINVAL; + goto bail; } /* No sense in reading the slots of nodes that don't exist @@ -977,31 +991,27 @@ static int o2hb_do_disk_heartbeat(struct o2hb_region *reg) ret = o2hb_read_slots(reg, highest_node + 1); if (ret < 0) { mlog_errno(ret); - return ret; + goto bail; } /* With an up to date view of the slots, we can check that no * other node has been improperly configured to heartbeat in * our slot. */ - if (!o2hb_check_last_timestamp(reg)) - mlog(ML_ERROR, "Device \"%s\": another node is heartbeating " - "in our slot!\n", reg->hr_dev_name); + own_slot_ok = o2hb_check_own_slot(reg); /* fill in the proper info for our next heartbeat */ o2hb_prepare_block(reg, reg->hr_generation); - /* And fire off the write. Note that we don't wait on this I/O - * until later. */ ret = o2hb_issue_node_write(reg, &write_wc); if (ret < 0) { mlog_errno(ret); - return ret; + goto bail; } i = -1; - while((i = find_next_bit(configured_nodes, O2NM_MAX_NODES, i + 1)) < O2NM_MAX_NODES) { - - change |= o2hb_check_slot(reg, ®->hr_slots[i]); + while((i = find_next_bit(configured_nodes, + O2NM_MAX_NODES, i + 1)) < O2NM_MAX_NODES) { + membership_change |= o2hb_check_slot(reg, ®->hr_slots[i]); } /* @@ -1016,18 +1026,39 @@ static int o2hb_do_disk_heartbeat(struct o2hb_region *reg) * disk */ mlog(ML_ERROR, "Write error %d on device \"%s\"\n", write_wc.wc_error, reg->hr_dev_name); - return write_wc.wc_error; + ret = write_wc.wc_error; + goto bail; } - o2hb_arm_write_timeout(reg); + /* Skip disarming the timeout if own slot has stale/bad data */ + if (own_slot_ok) { + o2hb_set_quorum_device(reg); + o2hb_arm_write_timeout(reg); + } +bail: /* let the person who launched us know when things are steady */ - if (!change && (atomic_read(®->hr_steady_iterations) != 0)) { - if (atomic_dec_and_test(®->hr_steady_iterations)) + if (atomic_read(®->hr_steady_iterations) != 0) { + if (!ret && own_slot_ok && !membership_change) { + if (atomic_dec_and_test(®->hr_steady_iterations)) + wake_up(&o2hb_steady_queue); + } + } + + if (atomic_read(®->hr_steady_iterations) != 0) { + if (atomic_dec_and_test(®->hr_unsteady_iterations)) { + printk(KERN_NOTICE "o2hb: Unable to stabilize " + "heartbeart on region %s (%s)\n", + config_item_name(®->hr_item), + reg->hr_dev_name); + atomic_set(®->hr_steady_iterations, 0); + reg->hr_aborted_start = 1; wake_up(&o2hb_steady_queue); + ret = -EIO; + } } - return 0; + return ret; } /* Subtract b from a, storing the result in a. a *must* have a larger @@ -1076,12 +1107,13 @@ static int o2hb_thread(void *data) mlog(ML_HEARTBEAT|ML_KTHREAD, "hb thread running\n"); - set_user_nice(current, -20); + set_user_nice(current, MIN_NICE); /* Pin node */ o2nm_depend_this_node(); - while (!kthread_should_stop() && !reg->hr_unclean_stop) { + while (!kthread_should_stop() && + !reg->hr_unclean_stop && !reg->hr_aborted_start) { /* We track the time spent inside * o2hb_do_disk_heartbeat so that we avoid more than * hr_timeout_ms between disk writes. On busy systems @@ -1089,10 +1121,7 @@ static int o2hb_thread(void *data) * likely to time itself out. */ do_gettimeofday(&before_hb); - i = 0; - do { - ret = o2hb_do_disk_heartbeat(reg); - } while (ret && ++i < 2); + ret = o2hb_do_disk_heartbeat(reg); do_gettimeofday(&after_hb); elapsed_msec = o2hb_elapsed_msecs(&before_hb, &after_hb); @@ -1103,7 +1132,8 @@ static int o2hb_thread(void *data) after_hb.tv_sec, (unsigned long) after_hb.tv_usec, elapsed_msec); - if (elapsed_msec < reg->hr_timeout_ms) { + if (!kthread_should_stop() && + elapsed_msec < reg->hr_timeout_ms) { /* the kthread api has blocked signals for us so no * need to record the return value. */ msleep_interruptible(reg->hr_timeout_ms - elapsed_msec); @@ -1120,20 +1150,20 @@ static int o2hb_thread(void *data) * to timeout on this region when we could just as easily * write a clear generation - thus indicating to them that * this node has left this region. - * - * XXX: Should we skip this on unclean_stop? */ - o2hb_prepare_block(reg, 0); - ret = o2hb_issue_node_write(reg, &write_wc); - if (ret == 0) { - o2hb_wait_on_io(reg, &write_wc); - } else { - mlog_errno(ret); + */ + if (!reg->hr_unclean_stop && !reg->hr_aborted_start) { + o2hb_prepare_block(reg, 0); + ret = o2hb_issue_node_write(reg, &write_wc); + if (ret == 0) + o2hb_wait_on_io(reg, &write_wc); + else + mlog_errno(ret); } /* Unpin node */ o2nm_undepend_this_node(); - mlog(ML_HEARTBEAT|ML_KTHREAD, "hb thread exiting\n"); + mlog(ML_HEARTBEAT|ML_KTHREAD, "o2hb thread exiting\n"); return 0; } @@ -1144,6 +1174,7 @@ static int o2hb_debug_open(struct inode *inode, struct file *file) struct o2hb_debug_buf *db = inode->i_private; struct o2hb_region *reg; unsigned long map[BITS_TO_LONGS(O2NM_MAX_NODES)]; + unsigned long lts; char *buf = NULL; int i = -1; int out = 0; @@ -1180,9 +1211,11 @@ static int o2hb_debug_open(struct inode *inode, struct file *file) case O2HB_DB_TYPE_REGION_ELAPSED_TIME: reg = (struct o2hb_region *)db->db_data; - out += snprintf(buf + out, PAGE_SIZE - out, "%u\n", - jiffies_to_msecs(jiffies - - reg->hr_last_timeout_start)); + lts = reg->hr_last_timeout_start; + /* If 0, it has never been set before */ + if (lts) + lts = jiffies_to_msecs(jiffies - lts); + out += snprintf(buf + out, PAGE_SIZE - out, "%lu\n", lts); goto done; case O2HB_DB_TYPE_REGION_PINNED: @@ -1412,8 +1445,9 @@ static void o2hb_region_release(struct config_item *item) struct page *page; struct o2hb_region *reg = to_o2hb_region(item); - if (reg->hr_tmp_block) - kfree(reg->hr_tmp_block); + mlog(ML_HEARTBEAT, "hb region release (%s)\n", reg->hr_dev_name); + + kfree(reg->hr_tmp_block); if (reg->hr_slot_data) { for (i = 0; i < reg->hr_num_pages; i++) { @@ -1427,8 +1461,7 @@ static void o2hb_region_release(struct config_item *item) if (reg->hr_bdev) blkdev_put(reg->hr_bdev, FMODE_READ|FMODE_WRITE); - if (reg->hr_slots) - kfree(reg->hr_slots); + kfree(reg->hr_slots); kfree(reg->hr_db_regnum); kfree(reg->hr_db_livenodes); @@ -1687,9 +1720,10 @@ static ssize_t o2hb_region_dev_write(struct o2hb_region *reg, long fd; int sectsize; char *p = (char *)page; - struct file *filp = NULL; - struct inode *inode = NULL; + struct fd f; + struct inode *inode; ssize_t ret = -EINVAL; + int live_threshold; if (reg->hr_bdev) goto out; @@ -1706,26 +1740,26 @@ static ssize_t o2hb_region_dev_write(struct o2hb_region *reg, if (fd < 0 || fd >= INT_MAX) goto out; - filp = fget(fd); - if (filp == NULL) + f = fdget(fd); + if (f.file == NULL) goto out; if (reg->hr_blocks == 0 || reg->hr_start_block == 0 || reg->hr_block_bytes == 0) - goto out; + goto out2; - inode = igrab(filp->f_mapping->host); + inode = igrab(f.file->f_mapping->host); if (inode == NULL) - goto out; + goto out2; if (!S_ISBLK(inode->i_mode)) - goto out; + goto out3; - reg->hr_bdev = I_BDEV(filp->f_mapping->host); + reg->hr_bdev = I_BDEV(f.file->f_mapping->host); ret = blkdev_get(reg->hr_bdev, FMODE_WRITE | FMODE_READ, NULL); if (ret) { reg->hr_bdev = NULL; - goto out; + goto out3; } inode = NULL; @@ -1737,7 +1771,7 @@ static ssize_t o2hb_region_dev_write(struct o2hb_region *reg, "blocksize %u incorrect for device, expected %d", reg->hr_block_bytes, sectsize); ret = -EINVAL; - goto out; + goto out3; } o2hb_init_region_params(reg); @@ -1751,13 +1785,13 @@ static ssize_t o2hb_region_dev_write(struct o2hb_region *reg, ret = o2hb_map_slot_data(reg); if (ret) { mlog_errno(ret); - goto out; + goto out3; } ret = o2hb_populate_slot_data(reg); if (ret) { mlog_errno(ret); - goto out; + goto out3; } INIT_DELAYED_WORK(®->hr_write_timeout_work, o2hb_write_timeout); @@ -1766,15 +1800,28 @@ static ssize_t o2hb_region_dev_write(struct o2hb_region *reg, * A node is considered live after it has beat LIVE_THRESHOLD * times. We're not steady until we've given them a chance * _after_ our first read. + * The default threshold is bare minimum so as to limit the delay + * during mounts. For global heartbeat, the threshold doubled for the + * first region. */ - atomic_set(®->hr_steady_iterations, O2HB_LIVE_THRESHOLD + 1); + live_threshold = O2HB_LIVE_THRESHOLD; + if (o2hb_global_heartbeat_active()) { + spin_lock(&o2hb_live_lock); + if (bitmap_weight(o2hb_region_bitmap, O2NM_MAX_REGIONS) == 1) + live_threshold <<= 1; + spin_unlock(&o2hb_live_lock); + } + ++live_threshold; + atomic_set(®->hr_steady_iterations, live_threshold); + /* unsteady_iterations is double the steady_iterations */ + atomic_set(®->hr_unsteady_iterations, (live_threshold << 1)); hb_task = kthread_run(o2hb_thread, reg, "o2hb-%s", reg->hr_item.ci_name); if (IS_ERR(hb_task)) { ret = PTR_ERR(hb_task); mlog_errno(ret); - goto out; + goto out3; } spin_lock(&o2hb_live_lock); @@ -1784,15 +1831,13 @@ static ssize_t o2hb_region_dev_write(struct o2hb_region *reg, ret = wait_event_interruptible(o2hb_steady_queue, atomic_read(®->hr_steady_iterations) == 0); if (ret) { - /* We got interrupted (hello ptrace!). Clean up */ - spin_lock(&o2hb_live_lock); - hb_task = reg->hr_task; - reg->hr_task = NULL; - spin_unlock(&o2hb_live_lock); + atomic_set(®->hr_steady_iterations, 0); + reg->hr_aborted_start = 1; + } - if (hb_task) - kthread_stop(hb_task); - goto out; + if (reg->hr_aborted_start) { + ret = -EIO; + goto out3; } /* Ok, we were woken. Make sure it wasn't by drop_item() */ @@ -1808,14 +1853,14 @@ static ssize_t o2hb_region_dev_write(struct o2hb_region *reg, ret = -EIO; if (hb_task && o2hb_global_heartbeat_active()) - printk(KERN_NOTICE "o2hb: Heartbeat started on region %s\n", - config_item_name(®->hr_item)); + printk(KERN_NOTICE "o2hb: Heartbeat started on region %s (%s)\n", + config_item_name(®->hr_item), reg->hr_dev_name); +out3: + iput(inode); +out2: + fdput(f); out: - if (filp) - fput(filp); - if (inode) - iput(inode); if (ret < 0) { if (reg->hr_bdev) { blkdev_put(reg->hr_bdev, FMODE_READ|FMODE_WRITE); @@ -2067,13 +2112,6 @@ static void o2hb_heartbeat_group_drop_item(struct config_group *group, /* stop the thread when the user removes the region dir */ spin_lock(&o2hb_live_lock); - if (o2hb_global_heartbeat_active()) { - clear_bit(reg->hr_region_num, o2hb_region_bitmap); - clear_bit(reg->hr_region_num, o2hb_live_region_bitmap); - if (test_bit(reg->hr_region_num, o2hb_quorum_region_bitmap)) - quorum_region = 1; - clear_bit(reg->hr_region_num, o2hb_quorum_region_bitmap); - } hb_task = reg->hr_task; reg->hr_task = NULL; reg->hr_item_dropped = 1; @@ -2082,19 +2120,30 @@ static void o2hb_heartbeat_group_drop_item(struct config_group *group, if (hb_task) kthread_stop(hb_task); + if (o2hb_global_heartbeat_active()) { + spin_lock(&o2hb_live_lock); + clear_bit(reg->hr_region_num, o2hb_region_bitmap); + clear_bit(reg->hr_region_num, o2hb_live_region_bitmap); + if (test_bit(reg->hr_region_num, o2hb_quorum_region_bitmap)) + quorum_region = 1; + clear_bit(reg->hr_region_num, o2hb_quorum_region_bitmap); + spin_unlock(&o2hb_live_lock); + printk(KERN_NOTICE "o2hb: Heartbeat %s on region %s (%s)\n", + ((atomic_read(®->hr_steady_iterations) == 0) ? + "stopped" : "start aborted"), config_item_name(item), + reg->hr_dev_name); + } + /* * If we're racing a dev_write(), we need to wake them. They will * check reg->hr_task */ if (atomic_read(®->hr_steady_iterations) != 0) { + reg->hr_aborted_start = 1; atomic_set(®->hr_steady_iterations, 0); wake_up(&o2hb_steady_queue); } - if (o2hb_global_heartbeat_active()) - printk(KERN_NOTICE "o2hb: Heartbeat stopped on region %s\n", - config_item_name(®->hr_item)); - config_item_put(item); if (!o2hb_global_heartbeat_active() || !quorum_region) @@ -2109,7 +2158,7 @@ static void o2hb_heartbeat_group_drop_item(struct config_group *group, if (!o2hb_dependent_users) goto unlock; - if (o2hb_pop_count(&o2hb_quorum_region_bitmap, + if (bitmap_weight(o2hb_quorum_region_bitmap, O2NM_MAX_REGIONS) <= O2HB_PIN_CUT_OFF) o2hb_region_pin(NULL); @@ -2198,7 +2247,7 @@ ssize_t o2hb_heartbeat_group_mode_store(struct o2hb_heartbeat_group *group, if (strnicmp(page, o2hb_heartbeat_mode_desc[i], len)) continue; - ret = o2hb_global_hearbeat_mode_set(i); + ret = o2hb_global_heartbeat_mode_set(i); if (!ret) printk(KERN_NOTICE "o2hb: Heartbeat mode set to %s\n", o2hb_heartbeat_mode_desc[i]); @@ -2231,7 +2280,7 @@ static struct configfs_attribute *o2hb_heartbeat_group_attrs[] = { NULL, }; -static struct configfs_item_operations o2hb_hearbeat_group_item_ops = { +static struct configfs_item_operations o2hb_heartbeat_group_item_ops = { .show_attribute = o2hb_heartbeat_group_show, .store_attribute = o2hb_heartbeat_group_store, }; @@ -2243,7 +2292,7 @@ static struct configfs_group_operations o2hb_heartbeat_group_group_ops = { static struct config_item_type o2hb_heartbeat_group_type = { .ct_group_ops = &o2hb_heartbeat_group_group_ops, - .ct_item_ops = &o2hb_hearbeat_group_item_ops, + .ct_item_ops = &o2hb_heartbeat_group_item_ops, .ct_attrs = o2hb_heartbeat_group_attrs, .ct_owner = THIS_MODULE, }; @@ -2316,6 +2365,9 @@ static int o2hb_region_pin(const char *region_uuid) assert_spin_locked(&o2hb_live_lock); list_for_each_entry(reg, &o2hb_all_regions, hr_all_item) { + if (reg->hr_item_dropped) + continue; + uuid = config_item_name(®->hr_item); /* local heartbeat */ @@ -2366,6 +2418,9 @@ static void o2hb_region_unpin(const char *region_uuid) assert_spin_locked(&o2hb_live_lock); list_for_each_entry(reg, &o2hb_all_regions, hr_all_item) { + if (reg->hr_item_dropped) + continue; + uuid = config_item_name(®->hr_item); if (region_uuid) { if (strcmp(region_uuid, uuid)) @@ -2403,7 +2458,7 @@ static int o2hb_region_inc_user(const char *region_uuid) if (o2hb_dependent_users > 1) goto unlock; - if (o2hb_pop_count(&o2hb_quorum_region_bitmap, + if (bitmap_weight(o2hb_quorum_region_bitmap, O2NM_MAX_REGIONS) <= O2HB_PIN_CUT_OFF) ret = o2hb_region_pin(NULL); @@ -2437,8 +2492,7 @@ unlock: int o2hb_register_callback(const char *region_uuid, struct o2hb_callback_func *hc) { - struct o2hb_callback_func *tmp; - struct list_head *iter; + struct o2hb_callback_func *f; struct o2hb_callback *hbcall; int ret; @@ -2461,10 +2515,9 @@ int o2hb_register_callback(const char *region_uuid, down_write(&o2hb_callback_sem); - list_for_each(iter, &hbcall->list) { - tmp = list_entry(iter, struct o2hb_callback_func, hc_item); - if (hc->hc_priority < tmp->hc_priority) { - list_add_tail(&hc->hc_item, iter); + list_for_each_entry(f, &hbcall->list, hc_item) { + if (hc->hc_priority < f->hc_priority) { + list_add_tail(&hc->hc_item, &f->hc_item); break; } } @@ -2581,6 +2634,9 @@ int o2hb_get_all_regions(char *region_uuids, u8 max_regions) p = region_uuids; list_for_each_entry(reg, &o2hb_all_regions, hr_all_item) { + if (reg->hr_item_dropped) + continue; + mlog(0, "Region: %s\n", config_item_name(®->hr_item)); if (numregs < max_regions) { memcpy(p, config_item_name(®->hr_item), diff --git a/fs/ocfs2/cluster/masklog.h b/fs/ocfs2/cluster/masklog.h index baa2b9ef7ee..2260fb9e650 100644 --- a/fs/ocfs2/cluster/masklog.h +++ b/fs/ocfs2/cluster/masklog.h @@ -199,7 +199,8 @@ extern struct mlog_bits mlog_and_bits, mlog_not_bits; #define mlog_errno(st) do { \ int _st = (st); \ if (_st != -ERESTARTSYS && _st != -EINTR && \ - _st != AOP_TRUNCATED_PAGE && _st != -ENOSPC) \ + _st != AOP_TRUNCATED_PAGE && _st != -ENOSPC && \ + _st != -EDQUOT) \ mlog(ML_ERROR, "status = %lld\n", (long long)_st); \ } while (0) diff --git a/fs/ocfs2/cluster/netdebug.c b/fs/ocfs2/cluster/netdebug.c index 3a5835904b3..73ba81928bc 100644 --- a/fs/ocfs2/cluster/netdebug.c +++ b/fs/ocfs2/cluster/netdebug.c @@ -47,6 +47,7 @@ #define SC_DEBUG_NAME "sock_containers" #define NST_DEBUG_NAME "send_tracking" #define STATS_DEBUG_NAME "stats" +#define NODES_DEBUG_NAME "connected_nodes" #define SHOW_SOCK_CONTAINERS 0 #define SHOW_SOCK_STATS 1 @@ -55,6 +56,7 @@ static struct dentry *o2net_dentry; static struct dentry *sc_dentry; static struct dentry *nst_dentry; static struct dentry *stats_dentry; +static struct dentry *nodes_dentry; static DEFINE_SPINLOCK(o2net_debug_lock); @@ -491,53 +493,87 @@ static const struct file_operations sc_seq_fops = { .release = sc_fop_release, }; -int o2net_debugfs_init(void) +static int o2net_fill_bitmap(char *buf, int len) { - o2net_dentry = debugfs_create_dir(O2NET_DEBUG_DIR, NULL); - if (!o2net_dentry) { - mlog_errno(-ENOMEM); - goto bail; - } + unsigned long map[BITS_TO_LONGS(O2NM_MAX_NODES)]; + int i = -1, out = 0; - nst_dentry = debugfs_create_file(NST_DEBUG_NAME, S_IFREG|S_IRUSR, - o2net_dentry, NULL, - &nst_seq_fops); - if (!nst_dentry) { - mlog_errno(-ENOMEM); - goto bail; - } + o2net_fill_node_map(map, sizeof(map)); - sc_dentry = debugfs_create_file(SC_DEBUG_NAME, S_IFREG|S_IRUSR, - o2net_dentry, NULL, - &sc_seq_fops); - if (!sc_dentry) { - mlog_errno(-ENOMEM); - goto bail; - } + while ((i = find_next_bit(map, O2NM_MAX_NODES, i + 1)) < O2NM_MAX_NODES) + out += snprintf(buf + out, PAGE_SIZE - out, "%d ", i); + out += snprintf(buf + out, PAGE_SIZE - out, "\n"); - stats_dentry = debugfs_create_file(STATS_DEBUG_NAME, S_IFREG|S_IRUSR, - o2net_dentry, NULL, - &stats_seq_fops); - if (!stats_dentry) { - mlog_errno(-ENOMEM); - goto bail; - } + return out; +} + +static int nodes_fop_open(struct inode *inode, struct file *file) +{ + char *buf; + + buf = kmalloc(PAGE_SIZE, GFP_KERNEL); + if (!buf) + return -ENOMEM; + + i_size_write(inode, o2net_fill_bitmap(buf, PAGE_SIZE)); + + file->private_data = buf; return 0; -bail: - debugfs_remove(stats_dentry); - debugfs_remove(sc_dentry); - debugfs_remove(nst_dentry); - debugfs_remove(o2net_dentry); - return -ENOMEM; } +static int o2net_debug_release(struct inode *inode, struct file *file) +{ + kfree(file->private_data); + return 0; +} + +static ssize_t o2net_debug_read(struct file *file, char __user *buf, + size_t nbytes, loff_t *ppos) +{ + return simple_read_from_buffer(buf, nbytes, ppos, file->private_data, + i_size_read(file->f_mapping->host)); +} + +static const struct file_operations nodes_fops = { + .open = nodes_fop_open, + .release = o2net_debug_release, + .read = o2net_debug_read, + .llseek = generic_file_llseek, +}; + void o2net_debugfs_exit(void) { + debugfs_remove(nodes_dentry); debugfs_remove(stats_dentry); debugfs_remove(sc_dentry); debugfs_remove(nst_dentry); debugfs_remove(o2net_dentry); } +int o2net_debugfs_init(void) +{ + umode_t mode = S_IFREG|S_IRUSR; + + o2net_dentry = debugfs_create_dir(O2NET_DEBUG_DIR, NULL); + if (o2net_dentry) + nst_dentry = debugfs_create_file(NST_DEBUG_NAME, mode, + o2net_dentry, NULL, &nst_seq_fops); + if (nst_dentry) + sc_dentry = debugfs_create_file(SC_DEBUG_NAME, mode, + o2net_dentry, NULL, &sc_seq_fops); + if (sc_dentry) + stats_dentry = debugfs_create_file(STATS_DEBUG_NAME, mode, + o2net_dentry, NULL, &stats_seq_fops); + if (stats_dentry) + nodes_dentry = debugfs_create_file(NODES_DEBUG_NAME, mode, + o2net_dentry, NULL, &nodes_fops); + if (nodes_dentry) + return 0; + + o2net_debugfs_exit(); + mlog_errno(-ENOMEM); + return -ENOMEM; +} + #endif /* CONFIG_DEBUG_FS */ diff --git a/fs/ocfs2/cluster/nodemanager.c b/fs/ocfs2/cluster/nodemanager.c index bb240647ca5..441c84e169e 100644 --- a/fs/ocfs2/cluster/nodemanager.c +++ b/fs/ocfs2/cluster/nodemanager.c @@ -29,7 +29,6 @@ #include "heartbeat.h" #include "masklog.h" #include "sys.h" -#include "ver.h" /* for now we operate under the assertion that there can be only one * cluster active at a time. Changing this will require trickling @@ -945,8 +944,6 @@ static int __init init_o2nm(void) { int ret = -1; - cluster_print_version(); - ret = o2hb_init(); if (ret) goto out; @@ -984,6 +981,7 @@ out: MODULE_AUTHOR("Oracle"); MODULE_LICENSE("GPL"); +MODULE_DESCRIPTION("OCFS2 cluster management"); module_init(init_o2nm) module_exit(exit_o2nm) diff --git a/fs/ocfs2/cluster/quorum.c b/fs/ocfs2/cluster/quorum.c index 8f9cea1597a..1ec141e758d 100644 --- a/fs/ocfs2/cluster/quorum.c +++ b/fs/ocfs2/cluster/quorum.c @@ -264,7 +264,7 @@ void o2quo_hb_still_up(u8 node) /* This is analogous to hb_up. as a node's connection comes up we delay the * quorum decision until we see it heartbeating. the hold will be droped in * hb_up or hb_down. it might be perpetuated by con_err until hb_down. if - * it's already heartbeating we we might be dropping a hold that conn_up got. + * it's already heartbeating we might be dropping a hold that conn_up got. * */ void o2quo_conn_up(u8 node) { @@ -327,5 +327,5 @@ void o2quo_exit(void) { struct o2quo_state *qs = &o2quo_state; - flush_work_sync(&qs->qs_work); + flush_work(&qs->qs_work); } diff --git a/fs/ocfs2/cluster/sys.c b/fs/ocfs2/cluster/sys.c index bc702dab5d1..b7f57271d49 100644 --- a/fs/ocfs2/cluster/sys.c +++ b/fs/ocfs2/cluster/sys.c @@ -41,7 +41,7 @@ static ssize_t version_show(struct kobject *kobj, struct kobj_attribute *attr, return snprintf(buf, PAGE_SIZE, "%u\n", O2NM_API_VERSION); } static struct kobj_attribute attr_version = - __ATTR(interface_revision, S_IFREG | S_IRUGO, version_show, NULL); + __ATTR(interface_revision, S_IRUGO, version_show, NULL); static struct attribute *o2cb_attrs[] = { &attr_version.attr, @@ -57,7 +57,6 @@ static struct kset *o2cb_kset; void o2cb_sys_shutdown(void) { mlog_sys_shutdown(); - sysfs_remove_link(NULL, "o2cb"); kset_unregister(o2cb_kset); } @@ -69,14 +68,6 @@ int o2cb_sys_init(void) if (!o2cb_kset) return -ENOMEM; - /* - * Create this symlink for backwards compatibility with old - * versions of ocfs2-tools which look for things in /sys/o2cb. - */ - ret = sysfs_create_link(NULL, &o2cb_kset->kobj, "o2cb"); - if (ret) - goto error; - ret = sysfs_create_group(&o2cb_kset->kobj, &o2cb_attr_group); if (ret) goto error; diff --git a/fs/ocfs2/cluster/tcp.c b/fs/ocfs2/cluster/tcp.c index db5ee4b4f47..681691bc233 100644 --- a/fs/ocfs2/cluster/tcp.c +++ b/fs/ocfs2/cluster/tcp.c @@ -59,6 +59,7 @@ #include <linux/idr.h> #include <linux/kref.h> #include <linux/net.h> +#include <linux/export.h> #include <net/tcp.h> #include <asm/uaccess.h> @@ -107,7 +108,7 @@ static struct rb_root o2net_handler_tree = RB_ROOT; static struct o2net_node o2net_nodes[O2NM_MAX_NODES]; /* XXX someday we'll need better accounting */ -static struct socket *o2net_listen_sock = NULL; +static struct socket *o2net_listen_sock; /* * listen work is only queued by the listening socket callbacks on the @@ -136,7 +137,7 @@ static int o2net_sys_err_translations[O2NET_ERR_MAX] = static void o2net_sc_connect_completed(struct work_struct *work); static void o2net_rx_until_empty(struct work_struct *work); static void o2net_shutdown_sc(struct work_struct *work); -static void o2net_listen_data_ready(struct sock *sk, int bytes); +static void o2net_listen_data_ready(struct sock *sk); static void o2net_sc_send_keep_req(struct work_struct *work); static void o2net_idle_timer(unsigned long data); static void o2net_sc_postpone_idle(struct o2net_sock_container *sc); @@ -261,17 +262,17 @@ static void o2net_update_recv_stats(struct o2net_sock_container *sc) #endif /* CONFIG_OCFS2_FS_STATS */ -static inline int o2net_reconnect_delay(void) +static inline unsigned int o2net_reconnect_delay(void) { return o2nm_single_cluster->cl_reconnect_delay_ms; } -static inline int o2net_keepalive_delay(void) +static inline unsigned int o2net_keepalive_delay(void) { return o2nm_single_cluster->cl_keepalive_delay_ms; } -static inline int o2net_idle_timeout(void) +static inline unsigned int o2net_idle_timeout(void) { return o2nm_single_cluster->cl_idle_timeout_ms; } @@ -303,28 +304,22 @@ static u8 o2net_num_from_nn(struct o2net_node *nn) static int o2net_prep_nsw(struct o2net_node *nn, struct o2net_status_wait *nsw) { - int ret = 0; - - do { - if (!idr_pre_get(&nn->nn_status_idr, GFP_ATOMIC)) { - ret = -EAGAIN; - break; - } - spin_lock(&nn->nn_lock); - ret = idr_get_new(&nn->nn_status_idr, nsw, &nsw->ns_id); - if (ret == 0) - list_add_tail(&nsw->ns_node_item, - &nn->nn_status_list); - spin_unlock(&nn->nn_lock); - } while (ret == -EAGAIN); + int ret; - if (ret == 0) { - init_waitqueue_head(&nsw->ns_wq); - nsw->ns_sys_status = O2NET_ERR_NONE; - nsw->ns_status = 0; + spin_lock(&nn->nn_lock); + ret = idr_alloc(&nn->nn_status_idr, nsw, 0, 0, GFP_ATOMIC); + if (ret >= 0) { + nsw->ns_id = ret; + list_add_tail(&nsw->ns_node_item, &nn->nn_status_list); } + spin_unlock(&nn->nn_lock); + if (ret < 0) + return ret; - return ret; + init_waitqueue_head(&nsw->ns_wq); + nsw->ns_sys_status = O2NET_ERR_NONE; + nsw->ns_status = 0; + return 0; } static void o2net_complete_nsw_locked(struct o2net_node *nn, @@ -411,6 +406,9 @@ static void sc_kref_release(struct kref *kref) sc->sc_node = NULL; o2net_debug_del_sc(sc); + + if (sc->sc_page) + __free_page(sc->sc_page); kfree(sc); } @@ -545,8 +543,9 @@ static void o2net_set_nn_state(struct o2net_node *nn, } if (was_valid && !valid) { - printk(KERN_NOTICE "o2net: no longer connected to " - SC_NODEF_FMT "\n", SC_NODEF_ARGS(old_sc)); + if (old_sc) + printk(KERN_NOTICE "o2net: No longer connected to " + SC_NODEF_FMT "\n", SC_NODEF_ARGS(old_sc)); o2net_complete_nodes_nsw(nn); } @@ -555,7 +554,7 @@ static void o2net_set_nn_state(struct o2net_node *nn, cancel_delayed_work(&nn->nn_connect_expired); printk(KERN_NOTICE "o2net: %s " SC_NODEF_FMT "\n", o2nm_this_node() > sc->sc_node->nd_num ? - "connected to" : "accepted connection from", + "Connected to" : "Accepted connection from", SC_NODEF_ARGS(sc)); } @@ -598,9 +597,9 @@ static void o2net_set_nn_state(struct o2net_node *nn, } /* see o2net_register_callbacks() */ -static void o2net_data_ready(struct sock *sk, int bytes) +static void o2net_data_ready(struct sock *sk) { - void (*ready)(struct sock *sk, int bytes); + void (*ready)(struct sock *sk); read_lock(&sk->sk_callback_lock); if (sk->sk_user_data) { @@ -614,7 +613,7 @@ static void o2net_data_ready(struct sock *sk, int bytes) } read_unlock(&sk->sk_callback_lock); - ready(sk, bytes); + ready(sk); } /* see o2net_register_callbacks() */ @@ -635,19 +634,19 @@ static void o2net_state_change(struct sock *sk) state_change = sc->sc_state_change; switch(sk->sk_state) { - /* ignore connecting sockets as they make progress */ - case TCP_SYN_SENT: - case TCP_SYN_RECV: - break; - case TCP_ESTABLISHED: - o2net_sc_queue_work(sc, &sc->sc_connect_work); - break; - default: - printk(KERN_INFO "o2net: connection to " SC_NODEF_FMT - " shutdown, state %d\n", - SC_NODEF_ARGS(sc), sk->sk_state); - o2net_sc_queue_work(sc, &sc->sc_shutdown_work); - break; + /* ignore connecting sockets as they make progress */ + case TCP_SYN_SENT: + case TCP_SYN_RECV: + break; + case TCP_ESTABLISHED: + o2net_sc_queue_work(sc, &sc->sc_connect_work); + break; + default: + printk(KERN_INFO "o2net: Connection to " SC_NODEF_FMT + " shutdown, state %d\n", + SC_NODEF_ARGS(sc), sk->sk_state); + o2net_sc_queue_work(sc, &sc->sc_shutdown_work); + break; } out: read_unlock(&sk->sk_callback_lock); @@ -767,32 +766,32 @@ static struct o2net_msg_handler * o2net_handler_tree_lookup(u32 msg_type, u32 key, struct rb_node ***ret_p, struct rb_node **ret_parent) { - struct rb_node **p = &o2net_handler_tree.rb_node; - struct rb_node *parent = NULL; + struct rb_node **p = &o2net_handler_tree.rb_node; + struct rb_node *parent = NULL; struct o2net_msg_handler *nmh, *ret = NULL; int cmp; - while (*p) { - parent = *p; - nmh = rb_entry(parent, struct o2net_msg_handler, nh_node); + while (*p) { + parent = *p; + nmh = rb_entry(parent, struct o2net_msg_handler, nh_node); cmp = o2net_handler_cmp(nmh, msg_type, key); - if (cmp < 0) - p = &(*p)->rb_left; - else if (cmp > 0) - p = &(*p)->rb_right; - else { + if (cmp < 0) + p = &(*p)->rb_left; + else if (cmp > 0) + p = &(*p)->rb_right; + else { ret = nmh; - break; + break; } - } + } - if (ret_p != NULL) - *ret_p = p; - if (ret_parent != NULL) - *ret_parent = parent; + if (ret_p != NULL) + *ret_p = p; + if (ret_parent != NULL) + *ret_parent = parent; - return ret; + return ret; } static void o2net_handler_kref_release(struct kref *kref) @@ -869,7 +868,7 @@ int o2net_register_handler(u32 msg_type, u32 key, u32 max_len, /* we've had some trouble with handlers seemingly vanishing. */ mlog_bug_on_msg(o2net_handler_tree_lookup(msg_type, key, &p, &parent) == NULL, - "couldn't find handler we *just* registerd " + "couldn't find handler we *just* registered " "for type %u key %08x\n", msg_type, key); } write_unlock(&o2net_handler_lock); @@ -917,57 +916,30 @@ static struct o2net_msg_handler *o2net_handler_get(u32 msg_type, u32 key) static int o2net_recv_tcp_msg(struct socket *sock, void *data, size_t len) { - int ret; - mm_segment_t oldfs; - struct kvec vec = { - .iov_len = len, - .iov_base = data, - }; - struct msghdr msg = { - .msg_iovlen = 1, - .msg_iov = (struct iovec *)&vec, - .msg_flags = MSG_DONTWAIT, - }; - - oldfs = get_fs(); - set_fs(get_ds()); - ret = sock_recvmsg(sock, &msg, len, msg.msg_flags); - set_fs(oldfs); - - return ret; + struct kvec vec = { .iov_len = len, .iov_base = data, }; + struct msghdr msg = { .msg_flags = MSG_DONTWAIT, }; + return kernel_recvmsg(sock, &msg, &vec, 1, len, msg.msg_flags); } static int o2net_send_tcp_msg(struct socket *sock, struct kvec *vec, size_t veclen, size_t total) { int ret; - mm_segment_t oldfs; - struct msghdr msg = { - .msg_iov = (struct iovec *)vec, - .msg_iovlen = veclen, - }; + struct msghdr msg; if (sock == NULL) { ret = -EINVAL; goto out; } - oldfs = get_fs(); - set_fs(get_ds()); - ret = sock_sendmsg(sock, &msg, total); - set_fs(oldfs); - if (ret != total) { - mlog(ML_ERROR, "sendmsg returned %d instead of %zu\n", ret, - total); - if (ret >= 0) - ret = -EPIPE; /* should be smarter, I bet */ - goto out; - } - - ret = 0; + ret = kernel_sendmsg(sock, &msg, vec, veclen, total); + if (likely(ret == total)) + return 0; + mlog(ML_ERROR, "sendmsg returned %d instead of %zu\n", ret, total); + if (ret >= 0) + ret = -EPIPE; /* should be smarter, I bet */ out: - if (ret < 0) - mlog(0, "returning error: %d\n", ret); + mlog(0, "returning error: %d\n", ret); return ret; } @@ -1034,6 +1006,25 @@ static int o2net_tx_can_proceed(struct o2net_node *nn, return ret; } +/* Get a map of all nodes to which this node is currently connected to */ +void o2net_fill_node_map(unsigned long *map, unsigned bytes) +{ + struct o2net_sock_container *sc; + int node, ret; + + BUG_ON(bytes < (BITS_TO_LONGS(O2NM_MAX_NODES) * sizeof(unsigned long))); + + memset(map, 0, bytes); + for (node = 0; node < O2NM_MAX_NODES; ++node) { + o2net_tx_can_proceed(o2net_nn_from_num(node), &sc, &ret); + if (!ret) { + set_bit(node, map); + sc_put(sc); + } + } +} +EXPORT_SYMBOL_GPL(o2net_fill_node_map); + int o2net_send_message_vec(u32 msg_type, u32 key, struct kvec *caller_vec, size_t caller_veclen, u8 target_node, int *status) { @@ -1145,10 +1136,8 @@ out: o2net_debug_del_nst(&nst); /* must be before dropping sc and node */ if (sc) sc_put(sc); - if (vec) - kfree(vec); - if (msg) - kfree(msg); + kfree(vec); + kfree(msg); o2net_complete_nsw(nn, &nsw, 0, 0, 0); return ret; } @@ -1284,11 +1273,11 @@ static int o2net_check_handshake(struct o2net_sock_container *sc) struct o2net_node *nn = o2net_nn_from_num(sc->sc_node->nd_num); if (hand->protocol_version != cpu_to_be64(O2NET_PROTOCOL_VERSION)) { - mlog(ML_NOTICE, SC_NODEF_FMT " advertised net protocol " - "version %llu but %llu is required, disconnecting\n", - SC_NODEF_ARGS(sc), - (unsigned long long)be64_to_cpu(hand->protocol_version), - O2NET_PROTOCOL_VERSION); + printk(KERN_NOTICE "o2net: " SC_NODEF_FMT " Advertised net " + "protocol version %llu but %llu is required. " + "Disconnecting.\n", SC_NODEF_ARGS(sc), + (unsigned long long)be64_to_cpu(hand->protocol_version), + O2NET_PROTOCOL_VERSION); /* don't bother reconnecting if its the wrong version. */ o2net_ensure_shutdown(nn, sc, -ENOTCONN); @@ -1302,33 +1291,33 @@ static int o2net_check_handshake(struct o2net_sock_container *sc) */ if (be32_to_cpu(hand->o2net_idle_timeout_ms) != o2net_idle_timeout()) { - mlog(ML_NOTICE, SC_NODEF_FMT " uses a network idle timeout of " - "%u ms, but we use %u ms locally. disconnecting\n", - SC_NODEF_ARGS(sc), - be32_to_cpu(hand->o2net_idle_timeout_ms), - o2net_idle_timeout()); + printk(KERN_NOTICE "o2net: " SC_NODEF_FMT " uses a network " + "idle timeout of %u ms, but we use %u ms locally. " + "Disconnecting.\n", SC_NODEF_ARGS(sc), + be32_to_cpu(hand->o2net_idle_timeout_ms), + o2net_idle_timeout()); o2net_ensure_shutdown(nn, sc, -ENOTCONN); return -1; } if (be32_to_cpu(hand->o2net_keepalive_delay_ms) != o2net_keepalive_delay()) { - mlog(ML_NOTICE, SC_NODEF_FMT " uses a keepalive delay of " - "%u ms, but we use %u ms locally. disconnecting\n", - SC_NODEF_ARGS(sc), - be32_to_cpu(hand->o2net_keepalive_delay_ms), - o2net_keepalive_delay()); + printk(KERN_NOTICE "o2net: " SC_NODEF_FMT " uses a keepalive " + "delay of %u ms, but we use %u ms locally. " + "Disconnecting.\n", SC_NODEF_ARGS(sc), + be32_to_cpu(hand->o2net_keepalive_delay_ms), + o2net_keepalive_delay()); o2net_ensure_shutdown(nn, sc, -ENOTCONN); return -1; } if (be32_to_cpu(hand->o2hb_heartbeat_timeout_ms) != O2HB_MAX_WRITE_TIMEOUT_MS) { - mlog(ML_NOTICE, SC_NODEF_FMT " uses a heartbeat timeout of " - "%u ms, but we use %u ms locally. disconnecting\n", - SC_NODEF_ARGS(sc), - be32_to_cpu(hand->o2hb_heartbeat_timeout_ms), - O2HB_MAX_WRITE_TIMEOUT_MS); + printk(KERN_NOTICE "o2net: " SC_NODEF_FMT " uses a heartbeat " + "timeout of %u ms, but we use %u ms locally. " + "Disconnecting.\n", SC_NODEF_ARGS(sc), + be32_to_cpu(hand->o2hb_heartbeat_timeout_ms), + O2HB_MAX_WRITE_TIMEOUT_MS); o2net_ensure_shutdown(nn, sc, -ENOTCONN); return -1; } @@ -1539,28 +1528,16 @@ static void o2net_idle_timer(unsigned long data) { struct o2net_sock_container *sc = (struct o2net_sock_container *)data; struct o2net_node *nn = o2net_nn_from_num(sc->sc_node->nd_num); - #ifdef CONFIG_DEBUG_FS - ktime_t now = ktime_get(); + unsigned long msecs = ktime_to_ms(ktime_get()) - + ktime_to_ms(sc->sc_tv_timer); +#else + unsigned long msecs = o2net_idle_timeout(); #endif - printk(KERN_NOTICE "o2net: connection to " SC_NODEF_FMT " has been idle for %u.%u " - "seconds, shutting it down.\n", SC_NODEF_ARGS(sc), - o2net_idle_timeout() / 1000, - o2net_idle_timeout() % 1000); - -#ifdef CONFIG_DEBUG_FS - mlog(ML_NOTICE, "Here are some times that might help debug the " - "situation: (Timer: %lld, Now %lld, DataReady %lld, Advance %lld-%lld, " - "Key 0x%08x, Func %u, FuncTime %lld-%lld)\n", - (long long)ktime_to_us(sc->sc_tv_timer), (long long)ktime_to_us(now), - (long long)ktime_to_us(sc->sc_tv_data_ready), - (long long)ktime_to_us(sc->sc_tv_advance_start), - (long long)ktime_to_us(sc->sc_tv_advance_stop), - sc->sc_msg_key, sc->sc_msg_type, - (long long)ktime_to_us(sc->sc_tv_func_start), - (long long)ktime_to_us(sc->sc_tv_func_stop)); -#endif + printk(KERN_NOTICE "o2net: Connection to " SC_NODEF_FMT " has been " + "idle for %lu.%lu secs, shutting it down.\n", SC_NODEF_ARGS(sc), + msecs / 1000, msecs % 1000); /* * Initialize the nn_timeout so that the next connection attempt @@ -1692,13 +1669,12 @@ static void o2net_start_connect(struct work_struct *work) ret = 0; out: - if (ret) { - mlog(ML_NOTICE, "connect attempt to " SC_NODEF_FMT " failed " - "with errno %d\n", SC_NODEF_ARGS(sc), ret); + if (ret && sc) { + printk(KERN_NOTICE "o2net: Connect attempt to " SC_NODEF_FMT + " failed with errno %d\n", SC_NODEF_ARGS(sc), ret); /* 0 err so that another will be queued and attempted * from set_nn_state */ - if (sc) - o2net_ensure_shutdown(nn, sc, 0); + o2net_ensure_shutdown(nn, sc, 0); } if (sc) sc_put(sc); @@ -1717,8 +1693,8 @@ static void o2net_connect_expired(struct work_struct *work) spin_lock(&nn->nn_lock); if (!nn->nn_sc_valid) { - mlog(ML_ERROR, "no connection established with node %u after " - "%u.%u seconds, giving up and returning errors.\n", + printk(KERN_NOTICE "o2net: No connection established with " + "node %u after %u.%u seconds, giving up.\n", o2net_num_from_nn(nn), o2net_idle_timeout() / 1000, o2net_idle_timeout() % 1000); @@ -1823,7 +1799,7 @@ int o2net_register_hb_callbacks(void) /* ------------------------------------------------------------ */ -static int o2net_accept_one(struct socket *sock) +static int o2net_accept_one(struct socket *sock, int *more) { int ret, slen; struct sockaddr_in sin; @@ -1834,6 +1810,7 @@ static int o2net_accept_one(struct socket *sock) struct o2net_node *nn; BUG_ON(sock == NULL); + *more = 0; ret = sock_create_lite(sock->sk->sk_family, sock->sk->sk_type, sock->sk->sk_protocol, &new_sock); if (ret) @@ -1845,6 +1822,7 @@ static int o2net_accept_one(struct socket *sock) if (ret < 0) goto out; + *more = 1; new_sock->sk->sk_allocation = GFP_ATOMIC; ret = o2net_set_nodelay(new_sock); @@ -1861,21 +1839,25 @@ static int o2net_accept_one(struct socket *sock) node = o2nm_get_node_by_ip(sin.sin_addr.s_addr); if (node == NULL) { - mlog(ML_NOTICE, "attempt to connect from unknown node at %pI4:%d\n", - &sin.sin_addr.s_addr, ntohs(sin.sin_port)); + printk(KERN_NOTICE "o2net: Attempt to connect from unknown " + "node at %pI4:%d\n", &sin.sin_addr.s_addr, + ntohs(sin.sin_port)); ret = -EINVAL; goto out; } if (o2nm_this_node() >= node->nd_num) { local_node = o2nm_get_node_by_num(o2nm_this_node()); - mlog(ML_NOTICE, "unexpected connect attempt seen at node '%s' (" - "%u, %pI4:%d) from node '%s' (%u, %pI4:%d)\n", - local_node->nd_name, local_node->nd_num, - &(local_node->nd_ipv4_address), - ntohs(local_node->nd_ipv4_port), - node->nd_name, node->nd_num, &sin.sin_addr.s_addr, - ntohs(sin.sin_port)); + if (local_node) + printk(KERN_NOTICE "o2net: Unexpected connect attempt " + "seen at node '%s' (%u, %pI4:%d) from " + "node '%s' (%u, %pI4:%d)\n", + local_node->nd_name, local_node->nd_num, + &(local_node->nd_ipv4_address), + ntohs(local_node->nd_ipv4_port), + node->nd_name, + node->nd_num, &sin.sin_addr.s_addr, + ntohs(sin.sin_port)); ret = -EINVAL; goto out; } @@ -1900,10 +1882,10 @@ static int o2net_accept_one(struct socket *sock) ret = 0; spin_unlock(&nn->nn_lock); if (ret) { - mlog(ML_NOTICE, "attempt to connect from node '%s' at " - "%pI4:%d but it already has an open connection\n", - node->nd_name, &sin.sin_addr.s_addr, - ntohs(sin.sin_port)); + printk(KERN_NOTICE "o2net: Attempt to connect from node '%s' " + "at %pI4:%d but it already has an open connection\n", + node->nd_name, &sin.sin_addr.s_addr, + ntohs(sin.sin_port)); goto out; } @@ -1939,16 +1921,41 @@ out: return ret; } +/* + * This function is invoked in response to one or more + * pending accepts at softIRQ level. We must drain the + * entire que before returning. + */ + static void o2net_accept_many(struct work_struct *work) { struct socket *sock = o2net_listen_sock; - while (o2net_accept_one(sock) == 0) + int more; + int err; + + /* + * It is critical to note that due to interrupt moderation + * at the network driver level, we can't assume to get a + * softIRQ for every single conn since tcp SYN packets + * can arrive back-to-back, and therefore many pending + * accepts may result in just 1 softIRQ. If we terminate + * the o2net_accept_one() loop upon seeing an err, what happens + * to the rest of the conns in the queue? If no new SYN + * arrives for hours, no softIRQ will be delivered, + * and the connections will just sit in the queue. + */ + + for (;;) { + err = o2net_accept_one(sock, &more); + if (!more) + break; cond_resched(); + } } -static void o2net_listen_data_ready(struct sock *sk, int bytes) +static void o2net_listen_data_ready(struct sock *sk) { - void (*ready)(struct sock *sk, int bytes); + void (*ready)(struct sock *sk); read_lock(&sk->sk_callback_lock); ready = sk->sk_user_data; @@ -1957,18 +1964,29 @@ static void o2net_listen_data_ready(struct sock *sk, int bytes) goto out; } - /* ->sk_data_ready is also called for a newly established child socket - * before it has been accepted and the acceptor has set up their - * data_ready.. we only want to queue listen work for our listening - * socket */ + /* This callback may called twice when a new connection + * is being established as a child socket inherits everything + * from a parent LISTEN socket, including the data_ready cb of + * the parent. This leads to a hazard. In o2net_accept_one() + * we are still initializing the child socket but have not + * changed the inherited data_ready callback yet when + * data starts arriving. + * We avoid this hazard by checking the state. + * For the listening socket, the state will be TCP_LISTEN; for the new + * socket, will be TCP_ESTABLISHED. Also, in this case, + * sk->sk_user_data is not a valid function pointer. + */ + if (sk->sk_state == TCP_LISTEN) { - mlog(ML_TCP, "bytes: %d\n", bytes); queue_work(o2net_wq, &o2net_listen_work); + } else { + ready = NULL; } out: read_unlock(&sk->sk_callback_lock); - ready(sk, bytes); + if (ready != NULL) + ready(sk); } static int o2net_open_listening_sock(__be32 addr, __be16 port) @@ -1983,7 +2001,7 @@ static int o2net_open_listening_sock(__be32 addr, __be16 port) ret = sock_create(PF_INET, SOCK_STREAM, IPPROTO_TCP, &sock); if (ret < 0) { - mlog(ML_ERROR, "unable to create socket, ret=%d\n", ret); + printk(KERN_ERR "o2net: Error %d while creating socket\n", ret); goto out; } @@ -1997,19 +2015,18 @@ static int o2net_open_listening_sock(__be32 addr, __be16 port) o2net_listen_sock = sock; INIT_WORK(&o2net_listen_work, o2net_accept_many); - sock->sk->sk_reuse = 1; + sock->sk->sk_reuse = SK_CAN_REUSE; ret = sock->ops->bind(sock, (struct sockaddr *)&sin, sizeof(sin)); if (ret < 0) { - mlog(ML_ERROR, "unable to bind socket at %pI4:%u, " - "ret=%d\n", &addr, ntohs(port), ret); + printk(KERN_ERR "o2net: Error %d while binding socket at " + "%pI4:%u\n", ret, &addr, ntohs(port)); goto out; } ret = sock->ops->listen(sock, 64); - if (ret < 0) { - mlog(ML_ERROR, "unable to listen on %pI4:%u, ret=%d\n", - &addr, ntohs(port), ret); - } + if (ret < 0) + printk(KERN_ERR "o2net: Error %d while listening on %pI4:%u\n", + ret, &addr, ntohs(port)); out: if (ret) { diff --git a/fs/ocfs2/cluster/tcp.h b/fs/ocfs2/cluster/tcp.h index fd6179eb26d..5bada2a69b5 100644 --- a/fs/ocfs2/cluster/tcp.h +++ b/fs/ocfs2/cluster/tcp.h @@ -106,6 +106,8 @@ int o2net_register_handler(u32 msg_type, u32 key, u32 max_len, struct list_head *unreg_list); void o2net_unregister_handler_list(struct list_head *list); +void o2net_fill_node_map(unsigned long *map, unsigned bytes); + struct o2nm_node; int o2net_register_hb_callbacks(void); void o2net_unregister_hb_callbacks(void); diff --git a/fs/ocfs2/cluster/tcp_internal.h b/fs/ocfs2/cluster/tcp_internal.h index 4cbcb65784a..dc024367110 100644 --- a/fs/ocfs2/cluster/tcp_internal.h +++ b/fs/ocfs2/cluster/tcp_internal.h @@ -165,7 +165,7 @@ struct o2net_sock_container { /* original handlers for the sockets */ void (*sc_state_change)(struct sock *sk); - void (*sc_data_ready)(struct sock *sk, int bytes); + void (*sc_data_ready)(struct sock *sk); u32 sc_msg_key; u16 sc_msg_type; diff --git a/fs/ocfs2/cluster/ver.c b/fs/ocfs2/cluster/ver.c deleted file mode 100644 index a56eee6abad..00000000000 --- a/fs/ocfs2/cluster/ver.c +++ /dev/null @@ -1,42 +0,0 @@ -/* -*- mode: c; c-basic-offset: 8; -*- - * vim: noexpandtab sw=8 ts=8 sts=0: - * - * ver.c - * - * version string - * - * Copyright (C) 2002, 2005 Oracle. All rights reserved. - * - * This program is free software; you can redistribute it and/or - * modify it under the terms of the GNU General Public - * License as published by the Free Software Foundation; either - * version 2 of the License, or (at your option) any later version. - * - * This program is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * General Public License for more details. - * - * You should have received a copy of the GNU General Public - * License along with this program; if not, write to the - * Free Software Foundation, Inc., 59 Temple Place - Suite 330, - * Boston, MA 021110-1307, USA. - */ - -#include <linux/module.h> -#include <linux/kernel.h> - -#include "ver.h" - -#define CLUSTER_BUILD_VERSION "1.5.0" - -#define VERSION_STR "OCFS2 Node Manager " CLUSTER_BUILD_VERSION - -void cluster_print_version(void) -{ - printk(KERN_INFO "%s\n", VERSION_STR); -} - -MODULE_DESCRIPTION(VERSION_STR); - -MODULE_VERSION(CLUSTER_BUILD_VERSION); diff --git a/fs/ocfs2/cluster/ver.h b/fs/ocfs2/cluster/ver.h deleted file mode 100644 index 32554c3382c..00000000000 --- a/fs/ocfs2/cluster/ver.h +++ /dev/null @@ -1,31 +0,0 @@ -/* -*- mode: c; c-basic-offset: 8; -*- - * vim: noexpandtab sw=8 ts=8 sts=0: - * - * ver.h - * - * Function prototypes - * - * Copyright (C) 2005 Oracle. All rights reserved. - * - * This program is free software; you can redistribute it and/or - * modify it under the terms of the GNU General Public - * License as published by the Free Software Foundation; either - * version 2 of the License, or (at your option) any later version. - * - * This program is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * General Public License for more details. - * - * You should have received a copy of the GNU General Public - * License along with this program; if not, write to the - * Free Software Foundation, Inc., 59 Temple Place - Suite 330, - * Boston, MA 021110-1307, USA. - */ - -#ifndef O2CLUSTER_VER_H -#define O2CLUSTER_VER_H - -void cluster_print_version(void); - -#endif /* O2CLUSTER_VER_H */ diff --git a/fs/ocfs2/dcache.c b/fs/ocfs2/dcache.c index e5ba3481833..e2e05a106be 100644 --- a/fs/ocfs2/dcache.c +++ b/fs/ocfs2/dcache.c @@ -37,7 +37,6 @@ #include "dlmglue.h" #include "file.h" #include "inode.h" -#include "super.h" #include "ocfs2_trace.h" void ocfs2_dentry_attach_gen(struct dentry *dentry) @@ -49,14 +48,13 @@ void ocfs2_dentry_attach_gen(struct dentry *dentry) } -static int ocfs2_dentry_revalidate(struct dentry *dentry, - struct nameidata *nd) +static int ocfs2_dentry_revalidate(struct dentry *dentry, unsigned int flags) { struct inode *inode; int ret = 0; /* if all else fails, just return false */ struct ocfs2_super *osb; - if (nd && nd->flags & LOOKUP_RCU) + if (flags & LOOKUP_RCU) return -ECHILD; inode = dentry->d_inode; @@ -71,9 +69,10 @@ static int ocfs2_dentry_revalidate(struct dentry *dentry, */ if (inode == NULL) { unsigned long gen = (unsigned long) dentry->d_fsdata; - unsigned long pgen = - OCFS2_I(dentry->d_parent->d_inode)->ip_dir_lock_gen; - + unsigned long pgen; + spin_lock(&dentry->d_lock); + pgen = OCFS2_I(dentry->d_parent->d_inode)->ip_dir_lock_gen; + spin_unlock(&dentry->d_lock); trace_ocfs2_dentry_revalidate_negative(dentry->d_name.len, dentry->d_name.name, pgen, gen); @@ -170,13 +169,10 @@ struct dentry *ocfs2_find_local_alias(struct inode *inode, u64 parent_blkno, int skip_unhashed) { - struct list_head *p; - struct dentry *dentry = NULL; + struct dentry *dentry; spin_lock(&inode->i_lock); - list_for_each(p, &inode->i_dentry) { - dentry = list_entry(p, struct dentry, d_alias); - + hlist_for_each_entry(dentry, &inode->i_dentry, d_alias) { spin_lock(&dentry->d_lock); if (ocfs2_match_dentry(dentry, parent_blkno, skip_unhashed)) { trace_ocfs2_find_local_alias(dentry->d_name.len, @@ -184,16 +180,13 @@ struct dentry *ocfs2_find_local_alias(struct inode *inode, dget_dlock(dentry); spin_unlock(&dentry->d_lock); - break; + spin_unlock(&inode->i_lock); + return dentry; } spin_unlock(&dentry->d_lock); - - dentry = NULL; } - spin_unlock(&inode->i_lock); - - return dentry; + return NULL; } DEFINE_SPINLOCK(dentry_attach_lock); @@ -352,52 +345,6 @@ out_attach: return ret; } -DEFINE_SPINLOCK(dentry_list_lock); - -/* We limit the number of dentry locks to drop in one go. We have - * this limit so that we don't starve other users of ocfs2_wq. */ -#define DL_INODE_DROP_COUNT 64 - -/* Drop inode references from dentry locks */ -static void __ocfs2_drop_dl_inodes(struct ocfs2_super *osb, int drop_count) -{ - struct ocfs2_dentry_lock *dl; - - spin_lock(&dentry_list_lock); - while (osb->dentry_lock_list && (drop_count < 0 || drop_count--)) { - dl = osb->dentry_lock_list; - osb->dentry_lock_list = dl->dl_next; - spin_unlock(&dentry_list_lock); - iput(dl->dl_inode); - kfree(dl); - spin_lock(&dentry_list_lock); - } - spin_unlock(&dentry_list_lock); -} - -void ocfs2_drop_dl_inodes(struct work_struct *work) -{ - struct ocfs2_super *osb = container_of(work, struct ocfs2_super, - dentry_lock_work); - - __ocfs2_drop_dl_inodes(osb, DL_INODE_DROP_COUNT); - /* - * Don't queue dropping if umount is in progress. We flush the - * list in ocfs2_dismount_volume - */ - spin_lock(&dentry_list_lock); - if (osb->dentry_lock_list && - !ocfs2_test_osb_flag(osb, OCFS2_OSB_DROP_DENTRY_LOCK_IMMED)) - queue_work(ocfs2_wq, &osb->dentry_lock_work); - spin_unlock(&dentry_list_lock); -} - -/* Flush the whole work queue */ -void ocfs2_drop_all_dl_inodes(struct ocfs2_super *osb) -{ - __ocfs2_drop_dl_inodes(osb, -1); -} - /* * ocfs2_dentry_iput() and friends. * @@ -422,24 +369,16 @@ void ocfs2_drop_all_dl_inodes(struct ocfs2_super *osb) static void ocfs2_drop_dentry_lock(struct ocfs2_super *osb, struct ocfs2_dentry_lock *dl) { + iput(dl->dl_inode); ocfs2_simple_drop_lockres(osb, &dl->dl_lockres); ocfs2_lock_res_free(&dl->dl_lockres); - - /* We leave dropping of inode reference to ocfs2_wq as that can - * possibly lead to inode deletion which gets tricky */ - spin_lock(&dentry_list_lock); - if (!osb->dentry_lock_list && - !ocfs2_test_osb_flag(osb, OCFS2_OSB_DROP_DENTRY_LOCK_IMMED)) - queue_work(ocfs2_wq, &osb->dentry_lock_work); - dl->dl_next = osb->dentry_lock_list; - osb->dentry_lock_list = dl; - spin_unlock(&dentry_list_lock); + kfree(dl); } void ocfs2_dentry_lock_put(struct ocfs2_super *osb, struct ocfs2_dentry_lock *dl) { - int unlock; + int unlock = 0; BUG_ON(dl->dl_count == 0); diff --git a/fs/ocfs2/dcache.h b/fs/ocfs2/dcache.h index b79eff70995..55f58892b15 100644 --- a/fs/ocfs2/dcache.h +++ b/fs/ocfs2/dcache.h @@ -29,13 +29,8 @@ extern const struct dentry_operations ocfs2_dentry_ops; struct ocfs2_dentry_lock { - /* Use count of dentry lock */ unsigned int dl_count; - union { - /* Linked list of dentry locks to release */ - struct ocfs2_dentry_lock *dl_next; - u64 dl_parent_blkno; - }; + u64 dl_parent_blkno; /* * The ocfs2_dentry_lock keeps an inode reference until @@ -49,14 +44,9 @@ struct ocfs2_dentry_lock { int ocfs2_dentry_attach_lock(struct dentry *dentry, struct inode *inode, u64 parent_blkno); -extern spinlock_t dentry_list_lock; - void ocfs2_dentry_lock_put(struct ocfs2_super *osb, struct ocfs2_dentry_lock *dl); -void ocfs2_drop_dl_inodes(struct work_struct *work); -void ocfs2_drop_all_dl_inodes(struct ocfs2_super *osb); - struct dentry *ocfs2_find_local_alias(struct inode *inode, u64 parent_blkno, int skip_unhashed); diff --git a/fs/ocfs2/dir.c b/fs/ocfs2/dir.c index 9fe5b8fd658..0717662b4ae 100644 --- a/fs/ocfs2/dir.c +++ b/fs/ocfs2/dir.c @@ -67,7 +67,6 @@ #define NAMEI_RA_CHUNKS 2 #define NAMEI_RA_BLOCKS 4 #define NAMEI_RA_SIZE (NAMEI_RA_CHUNKS * NAMEI_RA_BLOCKS) -#define NAMEI_RA_INDEX(c,b) (((c) * NAMEI_RA_BLOCKS) + (b)) static unsigned char ocfs2_filetype_table[] = { DT_UNKNOWN, DT_REG, DT_DIR, DT_CHR, DT_BLK, DT_FIFO, DT_SOCK, DT_LNK @@ -1184,8 +1183,7 @@ static int __ocfs2_delete_entry(handle_t *handle, struct inode *dir, if (pde) le16_add_cpu(&pde->rec_len, le16_to_cpu(de->rec_len)); - else - de->inode = 0; + de->inode = 0; dir->i_version++; ocfs2_journal_dirty(handle, bh); goto bail; @@ -1763,11 +1761,10 @@ bail: static int ocfs2_dir_foreach_blk_id(struct inode *inode, u64 *f_version, - loff_t *f_pos, void *priv, - filldir_t filldir, int *filldir_err) + struct dir_context *ctx) { - int ret, i, filldir_ret; - unsigned long offset = *f_pos; + int ret, i; + unsigned long offset = ctx->pos; struct buffer_head *di_bh = NULL; struct ocfs2_dinode *di; struct ocfs2_inline_data *data; @@ -1783,8 +1780,7 @@ static int ocfs2_dir_foreach_blk_id(struct inode *inode, di = (struct ocfs2_dinode *)di_bh->b_data; data = &di->id2.i_data; - while (*f_pos < i_size_read(inode)) { -revalidate: + while (ctx->pos < i_size_read(inode)) { /* If the dir block has changed since the last call to * readdir(2), then we might be pointing to an invalid * dirent right now. Scan from the start of the block @@ -1804,50 +1800,31 @@ revalidate: break; i += le16_to_cpu(de->rec_len); } - *f_pos = offset = i; + ctx->pos = offset = i; *f_version = inode->i_version; } - de = (struct ocfs2_dir_entry *) (data->id_data + *f_pos); - if (!ocfs2_check_dir_entry(inode, de, di_bh, *f_pos)) { + de = (struct ocfs2_dir_entry *) (data->id_data + ctx->pos); + if (!ocfs2_check_dir_entry(inode, de, di_bh, ctx->pos)) { /* On error, skip the f_pos to the end. */ - *f_pos = i_size_read(inode); - goto out; + ctx->pos = i_size_read(inode); + break; } offset += le16_to_cpu(de->rec_len); if (le64_to_cpu(de->inode)) { - /* We might block in the next section - * if the data destination is - * currently swapped out. So, use a - * version stamp to detect whether or - * not the directory has been modified - * during the copy operation. - */ - u64 version = *f_version; unsigned char d_type = DT_UNKNOWN; if (de->file_type < OCFS2_FT_MAX) d_type = ocfs2_filetype_table[de->file_type]; - filldir_ret = filldir(priv, de->name, - de->name_len, - *f_pos, - le64_to_cpu(de->inode), - d_type); - if (filldir_ret) { - if (filldir_err) - *filldir_err = filldir_ret; - break; - } - if (version != *f_version) - goto revalidate; + if (!dir_emit(ctx, de->name, de->name_len, + le64_to_cpu(de->inode), d_type)) + goto out; } - *f_pos += le16_to_cpu(de->rec_len); + ctx->pos += le16_to_cpu(de->rec_len); } - out: brelse(di_bh); - return 0; } @@ -1857,27 +1834,26 @@ out: */ static int ocfs2_dir_foreach_blk_el(struct inode *inode, u64 *f_version, - loff_t *f_pos, void *priv, - filldir_t filldir, int *filldir_err) + struct dir_context *ctx, + bool persist) { - int error = 0; unsigned long offset, blk, last_ra_blk = 0; - int i, stored; + int i; struct buffer_head * bh, * tmp; struct ocfs2_dir_entry * de; struct super_block * sb = inode->i_sb; unsigned int ra_sectors = 16; + int stored = 0; - stored = 0; bh = NULL; - offset = (*f_pos) & (sb->s_blocksize - 1); + offset = ctx->pos & (sb->s_blocksize - 1); - while (!error && !stored && *f_pos < i_size_read(inode)) { - blk = (*f_pos) >> sb->s_blocksize_bits; + while (ctx->pos < i_size_read(inode)) { + blk = ctx->pos >> sb->s_blocksize_bits; if (ocfs2_read_dir_block(inode, blk, &bh, 0)) { /* Skip the corrupt dirblock and keep trying */ - *f_pos += sb->s_blocksize - offset; + ctx->pos += sb->s_blocksize - offset; continue; } @@ -1899,7 +1875,6 @@ static int ocfs2_dir_foreach_blk_el(struct inode *inode, ra_sectors = 8; } -revalidate: /* If the dir block has changed since the last call to * readdir(2), then we might be pointing to an invalid * dirent right now. Scan from the start of the block @@ -1919,93 +1894,64 @@ revalidate: i += le16_to_cpu(de->rec_len); } offset = i; - *f_pos = ((*f_pos) & ~(sb->s_blocksize - 1)) + ctx->pos = (ctx->pos & ~(sb->s_blocksize - 1)) | offset; *f_version = inode->i_version; } - while (!error && *f_pos < i_size_read(inode) + while (ctx->pos < i_size_read(inode) && offset < sb->s_blocksize) { de = (struct ocfs2_dir_entry *) (bh->b_data + offset); if (!ocfs2_check_dir_entry(inode, de, bh, offset)) { /* On error, skip the f_pos to the next block. */ - *f_pos = ((*f_pos) | (sb->s_blocksize - 1)) + 1; + ctx->pos = (ctx->pos | (sb->s_blocksize - 1)) + 1; brelse(bh); - goto out; + continue; } - offset += le16_to_cpu(de->rec_len); if (le64_to_cpu(de->inode)) { - /* We might block in the next section - * if the data destination is - * currently swapped out. So, use a - * version stamp to detect whether or - * not the directory has been modified - * during the copy operation. - */ - unsigned long version = *f_version; unsigned char d_type = DT_UNKNOWN; if (de->file_type < OCFS2_FT_MAX) d_type = ocfs2_filetype_table[de->file_type]; - error = filldir(priv, de->name, + if (!dir_emit(ctx, de->name, de->name_len, - *f_pos, le64_to_cpu(de->inode), - d_type); - if (error) { - if (filldir_err) - *filldir_err = error; - break; + d_type)) { + brelse(bh); + return 0; } - if (version != *f_version) - goto revalidate; - stored ++; + stored++; } - *f_pos += le16_to_cpu(de->rec_len); + offset += le16_to_cpu(de->rec_len); + ctx->pos += le16_to_cpu(de->rec_len); } offset = 0; brelse(bh); bh = NULL; + if (!persist && stored) + break; } - - stored = 0; -out: - return stored; + return 0; } static int ocfs2_dir_foreach_blk(struct inode *inode, u64 *f_version, - loff_t *f_pos, void *priv, filldir_t filldir, - int *filldir_err) + struct dir_context *ctx, + bool persist) { if (OCFS2_I(inode)->ip_dyn_features & OCFS2_INLINE_DATA_FL) - return ocfs2_dir_foreach_blk_id(inode, f_version, f_pos, priv, - filldir, filldir_err); - - return ocfs2_dir_foreach_blk_el(inode, f_version, f_pos, priv, filldir, - filldir_err); + return ocfs2_dir_foreach_blk_id(inode, f_version, ctx); + return ocfs2_dir_foreach_blk_el(inode, f_version, ctx, persist); } /* * This is intended to be called from inside other kernel functions, * so we fake some arguments. */ -int ocfs2_dir_foreach(struct inode *inode, loff_t *f_pos, void *priv, - filldir_t filldir) +int ocfs2_dir_foreach(struct inode *inode, struct dir_context *ctx) { - int ret = 0, filldir_err = 0; u64 version = inode->i_version; - - while (*f_pos < i_size_read(inode)) { - ret = ocfs2_dir_foreach_blk(inode, &version, f_pos, priv, - filldir, &filldir_err); - if (ret || filldir_err) - break; - } - - if (ret > 0) - ret = -EIO; - + ocfs2_dir_foreach_blk(inode, &version, ctx, true); return 0; } @@ -2013,15 +1959,15 @@ int ocfs2_dir_foreach(struct inode *inode, loff_t *f_pos, void *priv, * ocfs2_readdir() * */ -int ocfs2_readdir(struct file * filp, void * dirent, filldir_t filldir) +int ocfs2_readdir(struct file *file, struct dir_context *ctx) { int error = 0; - struct inode *inode = filp->f_path.dentry->d_inode; + struct inode *inode = file_inode(file); int lock_level = 0; trace_ocfs2_readdir((unsigned long long)OCFS2_I(inode)->ip_blkno); - error = ocfs2_inode_lock_atime(inode, filp->f_vfsmnt, &lock_level); + error = ocfs2_inode_lock_atime(inode, file->f_path.mnt, &lock_level); if (lock_level && error >= 0) { /* We release EX lock which used to update atime * and get PR lock again to reduce contention @@ -2037,8 +1983,7 @@ int ocfs2_readdir(struct file * filp, void * dirent, filldir_t filldir) goto bail_nolock; } - error = ocfs2_dir_foreach_blk(inode, &filp->f_version, &filp->f_pos, - dirent, filldir, NULL); + error = ocfs2_dir_foreach_blk(inode, &file->f_version, ctx, false); ocfs2_inode_unlock(inode, lock_level); if (error) @@ -2122,6 +2067,7 @@ bail: } struct ocfs2_empty_dir_priv { + struct dir_context ctx; unsigned seen_dot; unsigned seen_dot_dot; unsigned seen_other; @@ -2206,10 +2152,9 @@ out: int ocfs2_empty_dir(struct inode *inode) { int ret; - loff_t start = 0; - struct ocfs2_empty_dir_priv priv; - - memset(&priv, 0, sizeof(priv)); + struct ocfs2_empty_dir_priv priv = { + .ctx.actor = ocfs2_empty_dir_filldir, + }; if (ocfs2_dir_indexed(inode)) { ret = ocfs2_empty_dir_dx(inode, &priv); @@ -2221,7 +2166,7 @@ int ocfs2_empty_dir(struct inode *inode) */ } - ret = ocfs2_dir_foreach(inode, &start, &priv, ocfs2_empty_dir_filldir); + ret = ocfs2_dir_foreach(inode, &priv.ctx); if (ret) mlog_errno(ret); @@ -2292,7 +2237,7 @@ static int ocfs2_fill_new_dir_id(struct ocfs2_super *osb, ocfs2_journal_dirty(handle, di_bh); i_size_write(inode, size); - inode->i_nlink = 2; + set_nlink(inode, 2); inode->i_blocks = ocfs2_inode_sector_count(inode); ret = ocfs2_mark_inode_dirty(handle, inode, di_bh); @@ -2354,7 +2299,7 @@ static int ocfs2_fill_new_dir_el(struct ocfs2_super *osb, ocfs2_journal_dirty(handle, new_bh); i_size_write(inode, inode->i_sb->s_blocksize); - inode->i_nlink = 2; + set_nlink(inode, 2); inode->i_blocks = ocfs2_inode_sector_count(inode); status = ocfs2_mark_inode_dirty(handle, inode, fe_bh); if (status < 0) { @@ -2404,7 +2349,7 @@ static int ocfs2_dx_dir_attach_index(struct ocfs2_super *osb, dx_root_bh = sb_getblk(osb->sb, dr_blkno); if (dx_root_bh == NULL) { - ret = -EIO; + ret = -ENOMEM; goto out; } ocfs2_set_new_buffer_uptodate(INODE_CACHE(dir), dx_root_bh); @@ -2477,7 +2422,7 @@ static int ocfs2_dx_dir_format_cluster(struct ocfs2_super *osb, for (i = 0; i < num_dx_leaves; i++) { bh = sb_getblk(osb->sb, start_blk + i); if (bh == NULL) { - ret = -EIO; + ret = -ENOMEM; goto out; } dx_leaves[i] = bh; @@ -2868,7 +2813,7 @@ static int ocfs2_expand_inline_dir(struct inode *dir, struct buffer_head *di_bh, bytes = blocks_wanted << sb->s_blocksize_bits; struct ocfs2_super *osb = OCFS2_SB(dir->i_sb); struct ocfs2_inode_info *oi = OCFS2_I(dir); - struct ocfs2_alloc_context *data_ac; + struct ocfs2_alloc_context *data_ac = NULL; struct ocfs2_alloc_context *meta_ac = NULL; struct buffer_head *dirdata_bh = NULL; struct buffer_head *dx_root_bh = NULL; @@ -2984,7 +2929,7 @@ static int ocfs2_expand_inline_dir(struct inode *dir, struct buffer_head *di_bh, blkno = ocfs2_clusters_to_blocks(dir->i_sb, bit_off); dirdata_bh = sb_getblk(sb, blkno); if (!dirdata_bh) { - ret = -EIO; + ret = -ENOMEM; mlog_errno(ret); goto out_commit; } @@ -3012,6 +2957,7 @@ static int ocfs2_expand_inline_dir(struct inode *dir, struct buffer_head *di_bh, ocfs2_init_dir_trailer(dir, dirdata_bh, i); } + ocfs2_update_inode_fsync_trans(handle, dir, 1); ocfs2_journal_dirty(handle, dirdata_bh); if (ocfs2_supports_indexed_dirs(osb) && !dx_inline) { @@ -3060,6 +3006,7 @@ static int ocfs2_expand_inline_dir(struct inode *dir, struct buffer_head *di_bh, di->i_size = cpu_to_le64(sb->s_blocksize); di->i_ctime = di->i_mtime = cpu_to_le64(dir->i_ctime.tv_sec); di->i_ctime_nsec = di->i_mtime_nsec = cpu_to_le32(dir->i_ctime.tv_nsec); + ocfs2_update_inode_fsync_trans(handle, dir, 1); /* * This should never fail as our extent list is empty and all @@ -3214,7 +3161,7 @@ static int ocfs2_do_extend_dir(struct super_block *sb, *new_bh = sb_getblk(sb, p_blkno); if (!*new_bh) { - status = -EIO; + status = -ENOMEM; mlog_errno(status); goto bail; } @@ -3339,7 +3286,7 @@ static int ocfs2_extend_dir(struct ocfs2_super *osb, if (ocfs2_dir_resv_allowed(osb)) data_ac->ac_resv = &OCFS2_I(dir)->ip_la_data_resv; - credits = ocfs2_calc_extend_credits(sb, el, 1); + credits = ocfs2_calc_extend_credits(sb, el); } else { spin_unlock(&OCFS2_I(dir)->ip_lock); credits = OCFS2_SIMPLE_DIR_EXTEND_CREDITS; @@ -3393,6 +3340,7 @@ do_extend: } else { de->rec_len = cpu_to_le16(sb->s_blocksize); } + ocfs2_update_inode_fsync_trans(handle, dir, 1); ocfs2_journal_dirty(handle, new_bh); dir_i_size += dir->i_sb->s_blocksize; @@ -3771,7 +3719,7 @@ static int ocfs2_dx_dir_rebalance_credits(struct ocfs2_super *osb, { int credits = ocfs2_clusters_to_blocks(osb->sb, 2); - credits += ocfs2_calc_extend_credits(osb->sb, &dx_root->dr_list, 1); + credits += ocfs2_calc_extend_credits(osb->sb, &dx_root->dr_list); credits += ocfs2_quota_trans_credits(osb->sb); return credits; } @@ -3951,6 +3899,7 @@ out_commit: dquot_free_space_nodirty(dir, ocfs2_clusters_to_bytes(dir->i_sb, 1)); + ocfs2_update_inode_fsync_trans(handle, dir, 1); ocfs2_commit_trans(osb, handle); out: @@ -4189,6 +4138,7 @@ static int ocfs2_expand_inline_dx_root(struct inode *dir, mlog_errno(ret); did_quota = 0; + ocfs2_update_inode_fsync_trans(handle, dir, 1); ocfs2_journal_dirty(handle, dx_root_bh); out_commit: @@ -4456,6 +4406,7 @@ static int ocfs2_dx_dir_remove_index(struct inode *dir, di->i_dyn_features = cpu_to_le16(OCFS2_I(dir)->ip_dyn_features); spin_unlock(&OCFS2_I(dir)->ip_lock); di->i_dx_root = cpu_to_le64(0ULL); + ocfs2_update_inode_fsync_trans(handle, dir, 1); ocfs2_journal_dirty(handle, di_bh); diff --git a/fs/ocfs2/dir.h b/fs/ocfs2/dir.h index e683f3deb64..f0344b75b14 100644 --- a/fs/ocfs2/dir.h +++ b/fs/ocfs2/dir.h @@ -92,9 +92,8 @@ int ocfs2_find_files_on_disk(const char *name, struct ocfs2_dir_lookup_result *res); int ocfs2_lookup_ino_from_name(struct inode *dir, const char *name, int namelen, u64 *blkno); -int ocfs2_readdir(struct file *filp, void *dirent, filldir_t filldir); -int ocfs2_dir_foreach(struct inode *inode, loff_t *f_pos, void *priv, - filldir_t filldir); +int ocfs2_readdir(struct file *file, struct dir_context *ctx); +int ocfs2_dir_foreach(struct inode *inode, struct dir_context *ctx); int ocfs2_prepare_dir_for_insert(struct ocfs2_super *osb, struct inode *dir, struct buffer_head *parent_fe_bh, diff --git a/fs/ocfs2/dlm/Makefile b/fs/ocfs2/dlm/Makefile index c8a044efbb1..bd1aab1f49a 100644 --- a/fs/ocfs2/dlm/Makefile +++ b/fs/ocfs2/dlm/Makefile @@ -3,5 +3,5 @@ ccflags-y := -Ifs/ocfs2 obj-$(CONFIG_OCFS2_FS_O2CB) += ocfs2_dlm.o ocfs2_dlm-objs := dlmdomain.o dlmdebug.o dlmthread.o dlmrecovery.o \ - dlmmaster.o dlmast.o dlmconvert.o dlmlock.o dlmunlock.o dlmver.o + dlmmaster.o dlmast.o dlmconvert.o dlmlock.o dlmunlock.o diff --git a/fs/ocfs2/dlm/dlmast.c b/fs/ocfs2/dlm/dlmast.c index 3a3ed4bb794..b46278f9ae4 100644 --- a/fs/ocfs2/dlm/dlmast.c +++ b/fs/ocfs2/dlm/dlmast.c @@ -292,8 +292,8 @@ int dlm_proxy_ast_handler(struct o2net_msg *msg, u32 len, void *data, struct dlm_lock *lock = NULL; struct dlm_proxy_ast *past = (struct dlm_proxy_ast *) msg->buf; char *name; - struct list_head *iter, *head=NULL; - u64 cookie; + struct list_head *head = NULL; + __be64 cookie; u32 flags; u8 node; @@ -373,8 +373,7 @@ int dlm_proxy_ast_handler(struct o2net_msg *msg, u32 len, void *data, /* try convert queue for both ast/bast */ head = &res->converting; lock = NULL; - list_for_each(iter, head) { - lock = list_entry (iter, struct dlm_lock, list); + list_for_each_entry(lock, head, list) { if (lock->ml.cookie == cookie) goto do_ast; } @@ -385,8 +384,7 @@ int dlm_proxy_ast_handler(struct o2net_msg *msg, u32 len, void *data, else head = &res->granted; - list_for_each(iter, head) { - lock = list_entry (iter, struct dlm_lock, list); + list_for_each_entry(lock, head, list) { if (lock->ml.cookie == cookie) goto do_ast; } diff --git a/fs/ocfs2/dlm/dlmcommon.h b/fs/ocfs2/dlm/dlmcommon.h index 4bdf7baee34..fae17c640df 100644 --- a/fs/ocfs2/dlm/dlmcommon.h +++ b/fs/ocfs2/dlm/dlmcommon.h @@ -108,7 +108,6 @@ static inline int dlm_is_recovery_lock(const char *lock_name, int name_len) struct dlm_recovery_ctxt { struct list_head resources; - struct list_head received; struct list_head node_data; u8 new_master; u8 dead_node; @@ -144,6 +143,7 @@ struct dlm_ctxt wait_queue_head_t dlm_join_events; unsigned long live_nodes_map[BITS_TO_LONGS(O2NM_MAX_NODES)]; unsigned long domain_map[BITS_TO_LONGS(O2NM_MAX_NODES)]; + unsigned long exit_domain_map[BITS_TO_LONGS(O2NM_MAX_NODES)]; unsigned long recovery_map[BITS_TO_LONGS(O2NM_MAX_NODES)]; struct dlm_recovery_ctxt reco; spinlock_t master_lock; @@ -331,6 +331,7 @@ struct dlm_lock_resource u16 state; char lvb[DLM_LVB_LEN]; unsigned int inflight_locks; + unsigned int inflight_assert_workers; unsigned long refmap[BITS_TO_LONGS(O2NM_MAX_NODES)]; }; @@ -401,6 +402,18 @@ static inline int dlm_lvb_is_empty(char *lvb) return 1; } +static inline char *dlm_list_in_text(enum dlm_lockres_list idx) +{ + if (idx == DLM_GRANTED_LIST) + return "granted"; + else if (idx == DLM_CONVERTING_LIST) + return "converting"; + else if (idx == DLM_BLOCKED_LIST) + return "blocked"; + else + return "unknown"; +} + static inline struct list_head * dlm_list_idx_to_ptr(struct dlm_lock_resource *res, enum dlm_lockres_list idx) { @@ -448,6 +461,7 @@ enum { DLM_FINALIZE_RECO_MSG = 518, DLM_QUERY_REGION = 519, DLM_QUERY_NODEINFO = 520, + DLM_BEGIN_EXIT_DOMAIN_MSG = 521, }; struct dlm_reco_node_data @@ -665,7 +679,7 @@ struct dlm_query_join_packet { }; union dlm_query_join_response { - u32 intval; + __be32 intval; struct dlm_query_join_packet packet; }; @@ -741,8 +755,8 @@ struct dlm_query_region { struct dlm_node_info { u8 ni_nodenum; u8 pad1; - u16 ni_ipv4_port; - u32 ni_ipv4_address; + __be16 ni_ipv4_port; + __be32 ni_ipv4_address; }; struct dlm_query_nodeinfo { @@ -845,8 +859,8 @@ void dlm_complete_recovery_thread(struct dlm_ctxt *dlm); void dlm_wait_for_recovery(struct dlm_ctxt *dlm); void dlm_kick_recovery_thread(struct dlm_ctxt *dlm); int dlm_is_node_dead(struct dlm_ctxt *dlm, u8 node); -int dlm_wait_for_node_death(struct dlm_ctxt *dlm, u8 node, int timeout); -int dlm_wait_for_node_recovery(struct dlm_ctxt *dlm, u8 node, int timeout); +void dlm_wait_for_node_death(struct dlm_ctxt *dlm, u8 node, int timeout); +void dlm_wait_for_node_recovery(struct dlm_ctxt *dlm, u8 node, int timeout); void dlm_put(struct dlm_ctxt *dlm); struct dlm_ctxt *dlm_grab(struct dlm_ctxt *dlm); @@ -863,9 +877,8 @@ static inline void dlm_lockres_get(struct dlm_lock_resource *res) kref_get(&res->refs); } void dlm_lockres_put(struct dlm_lock_resource *res); -void __dlm_unhash_lockres(struct dlm_lock_resource *res); -void __dlm_insert_lockres(struct dlm_ctxt *dlm, - struct dlm_lock_resource *res); +void __dlm_unhash_lockres(struct dlm_ctxt *dlm, struct dlm_lock_resource *res); +void __dlm_insert_lockres(struct dlm_ctxt *dlm, struct dlm_lock_resource *res); struct dlm_lock_resource * __dlm_lookup_lockres_full(struct dlm_ctxt *dlm, const char *name, unsigned int len, @@ -888,46 +901,18 @@ struct dlm_lock_resource *dlm_new_lockres(struct dlm_ctxt *dlm, const char *name, unsigned int namelen); -#define dlm_lockres_set_refmap_bit(bit,res) \ - __dlm_lockres_set_refmap_bit(bit,res,__FILE__,__LINE__) -#define dlm_lockres_clear_refmap_bit(bit,res) \ - __dlm_lockres_clear_refmap_bit(bit,res,__FILE__,__LINE__) - -static inline void __dlm_lockres_set_refmap_bit(int bit, - struct dlm_lock_resource *res, - const char *file, - int line) -{ - //printk("%s:%d:%.*s: setting bit %d\n", file, line, - // res->lockname.len, res->lockname.name, bit); - set_bit(bit, res->refmap); -} +void dlm_lockres_set_refmap_bit(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res, int bit); +void dlm_lockres_clear_refmap_bit(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res, int bit); -static inline void __dlm_lockres_clear_refmap_bit(int bit, - struct dlm_lock_resource *res, - const char *file, - int line) -{ - //printk("%s:%d:%.*s: clearing bit %d\n", file, line, - // res->lockname.len, res->lockname.name, bit); - clear_bit(bit, res->refmap); -} +void dlm_lockres_drop_inflight_ref(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res); +void dlm_lockres_grab_inflight_ref(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res); -void __dlm_lockres_drop_inflight_ref(struct dlm_ctxt *dlm, - struct dlm_lock_resource *res, - const char *file, - int line); -void __dlm_lockres_grab_inflight_ref(struct dlm_ctxt *dlm, - struct dlm_lock_resource *res, - int new_lockres, - const char *file, - int line); -#define dlm_lockres_drop_inflight_ref(d,r) \ - __dlm_lockres_drop_inflight_ref(d,r,__FILE__,__LINE__) -#define dlm_lockres_grab_inflight_ref(d,r) \ - __dlm_lockres_grab_inflight_ref(d,r,0,__FILE__,__LINE__) -#define dlm_lockres_grab_inflight_ref_new(d,r) \ - __dlm_lockres_grab_inflight_ref(d,r,1,__FILE__,__LINE__) +void __dlm_lockres_grab_inflight_worker(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res); void dlm_queue_ast(struct dlm_ctxt *dlm, struct dlm_lock *lock); void dlm_queue_bast(struct dlm_ctxt *dlm, struct dlm_lock *lock); @@ -1097,11 +1082,9 @@ static inline int dlm_lock_compatible(int existing, int request) static inline int dlm_lock_on_list(struct list_head *head, struct dlm_lock *lock) { - struct list_head *iter; struct dlm_lock *tmplock; - list_for_each(iter, head) { - tmplock = list_entry(iter, struct dlm_lock, list); + list_for_each_entry(tmplock, head, list) { if (tmplock == lock) return 1; } diff --git a/fs/ocfs2/dlm/dlmconvert.c b/fs/ocfs2/dlm/dlmconvert.c index 29a886d1e82..e36d63ff178 100644 --- a/fs/ocfs2/dlm/dlmconvert.c +++ b/fs/ocfs2/dlm/dlmconvert.c @@ -123,7 +123,6 @@ static enum dlm_status __dlmconvert_master(struct dlm_ctxt *dlm, int *kick_thread) { enum dlm_status status = DLM_NORMAL; - struct list_head *iter; struct dlm_lock *tmplock=NULL; assert_spin_locked(&res->spinlock); @@ -185,16 +184,14 @@ static enum dlm_status __dlmconvert_master(struct dlm_ctxt *dlm, /* upconvert from here on */ status = DLM_NORMAL; - list_for_each(iter, &res->granted) { - tmplock = list_entry(iter, struct dlm_lock, list); + list_for_each_entry(tmplock, &res->granted, list) { if (tmplock == lock) continue; if (!dlm_lock_compatible(tmplock->ml.type, type)) goto switch_queues; } - list_for_each(iter, &res->converting) { - tmplock = list_entry(iter, struct dlm_lock, list); + list_for_each_entry(tmplock, &res->converting, list) { if (!dlm_lock_compatible(tmplock->ml.type, type)) goto switch_queues; /* existing conversion requests take precedence */ @@ -424,8 +421,8 @@ int dlm_convert_lock_handler(struct o2net_msg *msg, u32 len, void *data, struct dlm_ctxt *dlm = data; struct dlm_convert_lock *cnv = (struct dlm_convert_lock *)msg->buf; struct dlm_lock_resource *res = NULL; - struct list_head *iter; struct dlm_lock *lock = NULL; + struct dlm_lock *tmp_lock; struct dlm_lockstatus *lksb; enum dlm_status status = DLM_NORMAL; u32 flags; @@ -471,14 +468,13 @@ int dlm_convert_lock_handler(struct o2net_msg *msg, u32 len, void *data, dlm_error(status); goto leave; } - list_for_each(iter, &res->granted) { - lock = list_entry(iter, struct dlm_lock, list); - if (lock->ml.cookie == cnv->cookie && - lock->ml.node == cnv->node_idx) { + list_for_each_entry(tmp_lock, &res->granted, list) { + if (tmp_lock->ml.cookie == cnv->cookie && + tmp_lock->ml.node == cnv->node_idx) { + lock = tmp_lock; dlm_lock_get(lock); break; } - lock = NULL; } spin_unlock(&res->spinlock); if (!lock) { diff --git a/fs/ocfs2/dlm/dlmdebug.c b/fs/ocfs2/dlm/dlmdebug.c index 04a32be0aeb..18f13c2e4a1 100644 --- a/fs/ocfs2/dlm/dlmdebug.c +++ b/fs/ocfs2/dlm/dlmdebug.c @@ -30,6 +30,7 @@ #include <linux/sysctl.h> #include <linux/spinlock.h> #include <linux/debugfs.h> +#include <linux/export.h> #include "cluster/heartbeat.h" #include "cluster/nodemanager.h" @@ -95,7 +96,6 @@ static void __dlm_print_lock(struct dlm_lock *lock) void __dlm_print_one_lock_resource(struct dlm_lock_resource *res) { - struct list_head *iter2; struct dlm_lock *lock; char buf[DLM_LOCKID_NAME_MAX]; @@ -117,18 +117,15 @@ void __dlm_print_one_lock_resource(struct dlm_lock_resource *res) res->inflight_locks, atomic_read(&res->asts_reserved)); dlm_print_lockres_refmap(res); printk(" granted queue:\n"); - list_for_each(iter2, &res->granted) { - lock = list_entry(iter2, struct dlm_lock, list); + list_for_each_entry(lock, &res->granted, list) { __dlm_print_lock(lock); } printk(" converting queue:\n"); - list_for_each(iter2, &res->converting) { - lock = list_entry(iter2, struct dlm_lock, list); + list_for_each_entry(lock, &res->converting, list) { __dlm_print_lock(lock); } printk(" blocked queue:\n"); - list_for_each(iter2, &res->blocked) { - lock = list_entry(iter2, struct dlm_lock, list); + list_for_each_entry(lock, &res->blocked, list) { __dlm_print_lock(lock); } } @@ -341,7 +338,7 @@ void dlm_print_one_mle(struct dlm_master_list_entry *mle) #ifdef CONFIG_DEBUG_FS -static struct dentry *dlm_debugfs_root = NULL; +static struct dentry *dlm_debugfs_root; #define DLM_DEBUGFS_DIR "o2dlm" #define DLM_DEBUGFS_DLM_STATE "dlm_state" @@ -445,7 +442,6 @@ static int debug_mle_print(struct dlm_ctxt *dlm, char *buf, int len) { struct dlm_master_list_entry *mle; struct hlist_head *bucket; - struct hlist_node *list; int i, out = 0; unsigned long total = 0, longest = 0, bucket_count = 0; @@ -455,9 +451,7 @@ static int debug_mle_print(struct dlm_ctxt *dlm, char *buf, int len) spin_lock(&dlm->master_lock); for (i = 0; i < DLM_HASH_BUCKETS; i++) { bucket = dlm_master_hash(dlm, i); - hlist_for_each(list, bucket) { - mle = hlist_entry(list, struct dlm_master_list_entry, - master_hash_node); + hlist_for_each_entry(mle, bucket, master_hash_node) { ++total; ++bucket_count; if (len - out < 200) @@ -756,6 +750,12 @@ static int debug_state_print(struct dlm_ctxt *dlm, char *buf, int len) buf + out, len - out); out += snprintf(buf + out, len - out, "\n"); + /* Exit Domain Map: xx xx xx */ + out += snprintf(buf + out, len - out, "Exit Domain Map: "); + out += stringify_nodemap(dlm->exit_domain_map, O2NM_MAX_NODES, + buf + out, len - out); + out += snprintf(buf + out, len - out, "\n"); + /* Live Map: xx xx xx */ out += snprintf(buf + out, len - out, "Live Map: "); out += stringify_nodemap(dlm->live_nodes_map, O2NM_MAX_NODES, diff --git a/fs/ocfs2/dlm/dlmdomain.c b/fs/ocfs2/dlm/dlmdomain.c index 7540a492eab..39efc5057a3 100644 --- a/fs/ocfs2/dlm/dlmdomain.c +++ b/fs/ocfs2/dlm/dlmdomain.c @@ -43,8 +43,6 @@ #include "dlmdomain.h" #include "dlmdebug.h" -#include "dlmver.h" - #define MLOG_MASK_PREFIX (ML_DLM|ML_DLM_DOMAIN) #include "cluster/masklog.h" @@ -132,10 +130,12 @@ static DECLARE_WAIT_QUEUE_HEAD(dlm_domain_events); * New in version 1.1: * - Message DLM_QUERY_REGION added to support global heartbeat * - Message DLM_QUERY_NODEINFO added to allow online node removes + * New in version 1.2: + * - Message DLM_BEGIN_EXIT_DOMAIN_MSG added to mark start of exit domain */ static const struct dlm_protocol_version dlm_protocol = { .pv_major = 1, - .pv_minor = 1, + .pv_minor = 2, }; #define DLM_DOMAIN_BACKOFF_MS 200 @@ -155,16 +155,18 @@ static int dlm_protocol_compare(struct dlm_protocol_version *existing, static void dlm_unregister_domain_handlers(struct dlm_ctxt *dlm); -void __dlm_unhash_lockres(struct dlm_lock_resource *lockres) +void __dlm_unhash_lockres(struct dlm_ctxt *dlm, struct dlm_lock_resource *res) { - if (!hlist_unhashed(&lockres->hash_node)) { - hlist_del_init(&lockres->hash_node); - dlm_lockres_put(lockres); - } + if (hlist_unhashed(&res->hash_node)) + return; + + mlog(0, "%s: Unhash res %.*s\n", dlm->name, res->lockname.len, + res->lockname.name); + hlist_del_init(&res->hash_node); + dlm_lockres_put(res); } -void __dlm_insert_lockres(struct dlm_ctxt *dlm, - struct dlm_lock_resource *res) +void __dlm_insert_lockres(struct dlm_ctxt *dlm, struct dlm_lock_resource *res) { struct hlist_head *bucket; struct qstr *q; @@ -178,6 +180,9 @@ void __dlm_insert_lockres(struct dlm_ctxt *dlm, dlm_lockres_get(res); hlist_add_head(&res->hash_node, bucket); + + mlog(0, "%s: Hash res %.*s\n", dlm->name, res->lockname.len, + res->lockname.name); } struct dlm_lock_resource * __dlm_lookup_lockres_full(struct dlm_ctxt *dlm, @@ -186,7 +191,7 @@ struct dlm_lock_resource * __dlm_lookup_lockres_full(struct dlm_ctxt *dlm, unsigned int hash) { struct hlist_head *bucket; - struct hlist_node *list; + struct dlm_lock_resource *res; mlog(0, "%.*s\n", len, name); @@ -194,9 +199,7 @@ struct dlm_lock_resource * __dlm_lookup_lockres_full(struct dlm_ctxt *dlm, bucket = dlm_lockres_hash(dlm, hash); - hlist_for_each(list, bucket) { - struct dlm_lock_resource *res = hlist_entry(list, - struct dlm_lock_resource, hash_node); + hlist_for_each_entry(res, bucket, hash_node) { if (res->lockname.name[0] != name[0]) continue; if (unlikely(res->lockname.len != len)) @@ -255,22 +258,19 @@ struct dlm_lock_resource * dlm_lookup_lockres(struct dlm_ctxt *dlm, static struct dlm_ctxt * __dlm_lookup_domain_full(const char *domain, int len) { - struct dlm_ctxt *tmp = NULL; - struct list_head *iter; + struct dlm_ctxt *tmp; assert_spin_locked(&dlm_domain_lock); /* tmp->name here is always NULL terminated, * but domain may not be! */ - list_for_each(iter, &dlm_domains) { - tmp = list_entry (iter, struct dlm_ctxt, list); + list_for_each_entry(tmp, &dlm_domains, list) { if (strlen(tmp->name) == len && memcmp(tmp->name, domain, len)==0) - break; - tmp = NULL; + return tmp; } - return tmp; + return NULL; } /* For null terminated domain strings ONLY */ @@ -312,9 +312,7 @@ static void dlm_free_ctxt_mem(struct dlm_ctxt *dlm) if (dlm->master_hash) dlm_free_pagevec((void **)dlm->master_hash, DLM_HASH_PAGES); - if (dlm->name) - kfree(dlm->name); - + kfree(dlm->name); kfree(dlm); } @@ -361,25 +359,22 @@ static void __dlm_get(struct dlm_ctxt *dlm) * you shouldn't trust your pointer. */ struct dlm_ctxt *dlm_grab(struct dlm_ctxt *dlm) { - struct list_head *iter; - struct dlm_ctxt *target = NULL; + struct dlm_ctxt *target; + struct dlm_ctxt *ret = NULL; spin_lock(&dlm_domain_lock); - list_for_each(iter, &dlm_domains) { - target = list_entry (iter, struct dlm_ctxt, list); - + list_for_each_entry(target, &dlm_domains, list) { if (target == dlm) { __dlm_get(target); + ret = target; break; } - - target = NULL; } spin_unlock(&dlm_domain_lock); - return target; + return ret; } int dlm_domain_fully_joined(struct dlm_ctxt *dlm) @@ -449,14 +444,18 @@ redo_bucket: dropped = dlm_empty_lockres(dlm, res); spin_lock(&res->spinlock); - __dlm_lockres_calc_usage(dlm, res); - iter = res->hash_node.next; + if (dropped) + __dlm_lockres_calc_usage(dlm, res); + else + iter = res->hash_node.next; spin_unlock(&res->spinlock); dlm_lockres_put(res); - if (dropped) + if (dropped) { + cond_resched_lock(&dlm->spinlock); goto redo_bucket; + } } cond_resched_lock(&dlm->spinlock); num += n; @@ -486,6 +485,28 @@ static int dlm_no_joining_node(struct dlm_ctxt *dlm) return ret; } +static int dlm_begin_exit_domain_handler(struct o2net_msg *msg, u32 len, + void *data, void **ret_data) +{ + struct dlm_ctxt *dlm = data; + unsigned int node; + struct dlm_exit_domain *exit_msg = (struct dlm_exit_domain *) msg->buf; + + if (!dlm_grab(dlm)) + return 0; + + node = exit_msg->node_idx; + mlog(0, "%s: Node %u sent a begin exit domain message\n", dlm->name, node); + + spin_lock(&dlm->spinlock); + set_bit(node, dlm->exit_domain_map); + spin_unlock(&dlm->spinlock); + + dlm_put(dlm); + + return 0; +} + static void dlm_mark_domain_leaving(struct dlm_ctxt *dlm) { /* Yikes, a double spinlock! I need domain_lock for the dlm @@ -511,17 +532,17 @@ again: static void __dlm_print_nodes(struct dlm_ctxt *dlm) { - int node = -1; + int node = -1, num = 0; assert_spin_locked(&dlm->spinlock); - printk(KERN_NOTICE "o2dlm: Nodes in domain %s: ", dlm->name); - + printk("( "); while ((node = find_next_bit(dlm->domain_map, O2NM_MAX_NODES, node + 1)) < O2NM_MAX_NODES) { printk("%d ", node); + ++num; } - printk("\n"); + printk(") %u nodes\n", num); } static int dlm_exit_domain_handler(struct o2net_msg *msg, u32 len, void *data, @@ -538,10 +559,10 @@ static int dlm_exit_domain_handler(struct o2net_msg *msg, u32 len, void *data, node = exit_msg->node_idx; - printk(KERN_NOTICE "o2dlm: Node %u leaves domain %s\n", node, dlm->name); - spin_lock(&dlm->spinlock); clear_bit(node, dlm->domain_map); + clear_bit(node, dlm->exit_domain_map); + printk(KERN_NOTICE "o2dlm: Node %u leaves domain %s ", node, dlm->name); __dlm_print_nodes(dlm); /* notify anything attached to the heartbeat events */ @@ -554,29 +575,56 @@ static int dlm_exit_domain_handler(struct o2net_msg *msg, u32 len, void *data, return 0; } -static int dlm_send_one_domain_exit(struct dlm_ctxt *dlm, +static int dlm_send_one_domain_exit(struct dlm_ctxt *dlm, u32 msg_type, unsigned int node) { int status; struct dlm_exit_domain leave_msg; - mlog(0, "Asking node %u if we can leave the domain %s me = %u\n", - node, dlm->name, dlm->node_num); + mlog(0, "%s: Sending domain exit message %u to node %u\n", dlm->name, + msg_type, node); memset(&leave_msg, 0, sizeof(leave_msg)); leave_msg.node_idx = dlm->node_num; - status = o2net_send_message(DLM_EXIT_DOMAIN_MSG, dlm->key, - &leave_msg, sizeof(leave_msg), node, - NULL); + status = o2net_send_message(msg_type, dlm->key, &leave_msg, + sizeof(leave_msg), node, NULL); if (status < 0) - mlog(ML_ERROR, "Error %d when sending message %u (key 0x%x) to " - "node %u\n", status, DLM_EXIT_DOMAIN_MSG, dlm->key, node); - mlog(0, "status return %d from o2net_send_message\n", status); + mlog(ML_ERROR, "Error %d sending domain exit message %u " + "to node %u on domain %s\n", status, msg_type, node, + dlm->name); return status; } +static void dlm_begin_exit_domain(struct dlm_ctxt *dlm) +{ + int node = -1; + + /* Support for begin exit domain was added in 1.2 */ + if (dlm->dlm_locking_proto.pv_major == 1 && + dlm->dlm_locking_proto.pv_minor < 2) + return; + + /* + * Unlike DLM_EXIT_DOMAIN_MSG, DLM_BEGIN_EXIT_DOMAIN_MSG is purely + * informational. Meaning if a node does not receive the message, + * so be it. + */ + spin_lock(&dlm->spinlock); + while (1) { + node = find_next_bit(dlm->domain_map, O2NM_MAX_NODES, node + 1); + if (node >= O2NM_MAX_NODES) + break; + if (node == dlm->node_num) + continue; + + spin_unlock(&dlm->spinlock); + dlm_send_one_domain_exit(dlm, DLM_BEGIN_EXIT_DOMAIN_MSG, node); + spin_lock(&dlm->spinlock); + } + spin_unlock(&dlm->spinlock); +} static void dlm_leave_domain(struct dlm_ctxt *dlm) { @@ -602,7 +650,8 @@ static void dlm_leave_domain(struct dlm_ctxt *dlm) clear_node = 1; - status = dlm_send_one_domain_exit(dlm, node); + status = dlm_send_one_domain_exit(dlm, DLM_EXIT_DOMAIN_MSG, + node); if (status < 0 && status != -ENOPROTOOPT && status != -ENOTCONN) { @@ -677,6 +726,7 @@ void dlm_unregister_domain(struct dlm_ctxt *dlm) if (leave) { mlog(0, "shutting down domain %s\n", dlm->name); + dlm_begin_exit_domain(dlm); /* We changed dlm state, notify the thread */ dlm_kick_thread(dlm, NULL); @@ -697,6 +747,7 @@ void dlm_unregister_domain(struct dlm_ctxt *dlm) dlm_mark_domain_leaving(dlm); dlm_leave_domain(dlm); + printk(KERN_NOTICE "o2dlm: Leaving domain %s\n", dlm->name); dlm_force_free_mles(dlm); dlm_complete_dlm_shutdown(dlm); } @@ -755,7 +806,7 @@ static void dlm_query_join_packet_to_wire(struct dlm_query_join_packet *packet, union dlm_query_join_response response; response.packet = *packet; - *wire = cpu_to_be32(response.intval); + *wire = be32_to_cpu(response.intval); } static void dlm_query_join_wire_to_packet(u32 wire, @@ -908,10 +959,19 @@ static int dlm_assert_joined_handler(struct o2net_msg *msg, u32 len, void *data, * domain. Set him in the map and clean up our * leftover join state. */ BUG_ON(dlm->joining_node != assert->node_idx); + + if (dlm->reco.state & DLM_RECO_STATE_ACTIVE) { + mlog(0, "dlm recovery is ongoing, disallow join\n"); + spin_unlock(&dlm->spinlock); + spin_unlock(&dlm_domain_lock); + return -EAGAIN; + } + set_bit(assert->node_idx, dlm->domain_map); + clear_bit(assert->node_idx, dlm->exit_domain_map); __dlm_set_joining_node(dlm, DLM_LOCK_RES_OWNER_UNKNOWN); - printk(KERN_NOTICE "o2dlm: Node %u joins domain %s\n", + printk(KERN_NOTICE "o2dlm: Node %u joins domain %s ", assert->node_idx, dlm->name); __dlm_print_nodes(dlm); @@ -1071,7 +1131,6 @@ static int dlm_query_region_handler(struct o2net_msg *msg, u32 len, struct dlm_ctxt *dlm = NULL; char *local = NULL; int status = 0; - int locked = 0; qr = (struct dlm_query_region *) msg->buf; @@ -1080,10 +1139,8 @@ static int dlm_query_region_handler(struct o2net_msg *msg, u32 len, /* buffer used in dlm_mast_regions() */ local = kmalloc(sizeof(qr->qr_regions), GFP_KERNEL); - if (!local) { - status = -ENOMEM; - goto bail; - } + if (!local) + return -ENOMEM; status = -EINVAL; @@ -1092,16 +1149,15 @@ static int dlm_query_region_handler(struct o2net_msg *msg, u32 len, if (!dlm) { mlog(ML_ERROR, "Node %d queried hb regions on domain %s " "before join domain\n", qr->qr_node, qr->qr_domain); - goto bail; + goto out_domain_lock; } spin_lock(&dlm->spinlock); - locked = 1; if (dlm->joining_node != qr->qr_node) { mlog(ML_ERROR, "Node %d queried hb regions on domain %s " "but joining node is %d\n", qr->qr_node, qr->qr_domain, dlm->joining_node); - goto bail; + goto out_dlm_lock; } /* Support for global heartbeat was added in 1.1 */ @@ -1111,14 +1167,15 @@ static int dlm_query_region_handler(struct o2net_msg *msg, u32 len, "but active dlm protocol is %d.%d\n", qr->qr_node, qr->qr_domain, dlm->dlm_locking_proto.pv_major, dlm->dlm_locking_proto.pv_minor); - goto bail; + goto out_dlm_lock; } status = dlm_match_regions(dlm, qr, local, sizeof(qr->qr_regions)); -bail: - if (locked) - spin_unlock(&dlm->spinlock); +out_dlm_lock: + spin_unlock(&dlm->spinlock); + +out_domain_lock: spin_unlock(&dlm_domain_lock); kfree(local); @@ -1468,6 +1525,7 @@ static int dlm_send_one_join_assert(struct dlm_ctxt *dlm, unsigned int node) { int status; + int ret; struct dlm_assert_joined assert_msg; mlog(0, "Sending join assert to node %u\n", node); @@ -1479,11 +1537,13 @@ static int dlm_send_one_join_assert(struct dlm_ctxt *dlm, status = o2net_send_message(DLM_ASSERT_JOINED_MSG, DLM_MOD_KEY, &assert_msg, sizeof(assert_msg), node, - NULL); + &ret); if (status < 0) mlog(ML_ERROR, "Error %d when sending message %u (key 0x%x) to " "node %u\n", status, DLM_ASSERT_JOINED_MSG, DLM_MOD_KEY, node); + else + status = ret; return status; } @@ -1614,7 +1674,8 @@ static int dlm_try_to_join_domain(struct dlm_ctxt *dlm) spin_unlock(&dlm->spinlock); /* Support for global heartbeat and node info was added in 1.1 */ - if (dlm_protocol.pv_major > 1 || dlm_protocol.pv_minor > 0) { + if (dlm->dlm_locking_proto.pv_major > 1 || + dlm->dlm_locking_proto.pv_minor > 0) { status = dlm_send_nodeinfo(dlm, ctxt->yes_resp_map); if (status) { mlog_errno(status); @@ -1641,8 +1702,10 @@ static int dlm_try_to_join_domain(struct dlm_ctxt *dlm) bail: spin_lock(&dlm->spinlock); __dlm_set_joining_node(dlm, DLM_LOCK_RES_OWNER_UNKNOWN); - if (!status) + if (!status) { + printk(KERN_NOTICE "o2dlm: Joining domain %s ", dlm->name); __dlm_print_nodes(dlm); + } spin_unlock(&dlm->spinlock); if (ctxt) { @@ -1792,6 +1855,13 @@ static int dlm_register_domain_handlers(struct dlm_ctxt *dlm) if (status) goto bail; + status = o2net_register_handler(DLM_BEGIN_EXIT_DOMAIN_MSG, dlm->key, + sizeof(struct dlm_exit_domain), + dlm_begin_exit_domain_handler, + dlm, NULL, &dlm->dlm_domain_handlers); + if (status) + goto bail; + bail: if (status) dlm_unregister_domain_handlers(dlm); @@ -1815,19 +1885,19 @@ static int dlm_join_domain(struct dlm_ctxt *dlm) goto bail; } - status = dlm_debug_init(dlm); + status = dlm_launch_thread(dlm); if (status < 0) { mlog_errno(status); goto bail; } - status = dlm_launch_thread(dlm); + status = dlm_launch_recovery_thread(dlm); if (status < 0) { mlog_errno(status); goto bail; } - status = dlm_launch_recovery_thread(dlm); + status = dlm_debug_init(dlm); if (status < 0) { mlog_errno(status); goto bail; @@ -1964,7 +2034,6 @@ static struct dlm_ctxt *dlm_alloc_ctxt(const char *domain, INIT_LIST_HEAD(&dlm->list); INIT_LIST_HEAD(&dlm->dirty_list); INIT_LIST_HEAD(&dlm->reco.resources); - INIT_LIST_HEAD(&dlm->reco.received); INIT_LIST_HEAD(&dlm->reco.node_data); INIT_LIST_HEAD(&dlm->purge_list); INIT_LIST_HEAD(&dlm->dlm_domain_handlers); @@ -2064,13 +2133,6 @@ struct dlm_ctxt * dlm_register_domain(const char *domain, goto leave; } - if (!o2hb_check_local_node_heartbeating()) { - mlog(ML_ERROR, "the local node has not been configured, or is " - "not heartbeating\n"); - ret = -EPROTO; - goto leave; - } - mlog(0, "register called for domain \"%s\"\n", domain); retry: @@ -2231,13 +2293,10 @@ static DECLARE_RWSEM(dlm_callback_sem); void dlm_fire_domain_eviction_callbacks(struct dlm_ctxt *dlm, int node_num) { - struct list_head *iter; struct dlm_eviction_cb *cb; down_read(&dlm_callback_sem); - list_for_each(iter, &dlm->dlm_eviction_callbacks) { - cb = list_entry(iter, struct dlm_eviction_cb, ec_item); - + list_for_each_entry(cb, &dlm->dlm_eviction_callbacks, ec_item) { cb->ec_func(node_num, cb->ec_data); } up_read(&dlm_callback_sem); @@ -2274,8 +2333,6 @@ static int __init dlm_init(void) { int status; - dlm_print_version(); - status = dlm_init_mle_cache(); if (status) { mlog(ML_ERROR, "Could not create o2dlm_mle slabcache\n"); @@ -2325,6 +2382,7 @@ static void __exit dlm_exit (void) MODULE_AUTHOR("Oracle"); MODULE_LICENSE("GPL"); +MODULE_DESCRIPTION("OCFS2 Distributed Lock Management"); module_init(dlm_init); module_exit(dlm_exit); diff --git a/fs/ocfs2/dlm/dlmlock.c b/fs/ocfs2/dlm/dlmlock.c index 8d39e0fd66f..66c2a491f68 100644 --- a/fs/ocfs2/dlm/dlmlock.c +++ b/fs/ocfs2/dlm/dlmlock.c @@ -52,7 +52,7 @@ #define MLOG_MASK_PREFIX ML_DLM #include "cluster/masklog.h" -static struct kmem_cache *dlm_lock_cache = NULL; +static struct kmem_cache *dlm_lock_cache; static DEFINE_SPINLOCK(dlm_cookie_lock); static u64 dlm_next_cookie = 1; @@ -91,19 +91,14 @@ void dlm_destroy_lock_cache(void) static int dlm_can_grant_new_lock(struct dlm_lock_resource *res, struct dlm_lock *lock) { - struct list_head *iter; struct dlm_lock *tmplock; - list_for_each(iter, &res->granted) { - tmplock = list_entry(iter, struct dlm_lock, list); - + list_for_each_entry(tmplock, &res->granted, list) { if (!dlm_lock_compatible(tmplock->ml.type, lock->ml.type)) return 0; } - list_for_each(iter, &res->converting) { - tmplock = list_entry(iter, struct dlm_lock, list); - + list_for_each_entry(tmplock, &res->converting, list) { if (!dlm_lock_compatible(tmplock->ml.type, lock->ml.type)) return 0; if (!dlm_lock_compatible(tmplock->ml.convert_type, @@ -178,15 +173,12 @@ static enum dlm_status dlmlock_master(struct dlm_ctxt *dlm, lock->ml.node); } } else { + status = DLM_NORMAL; dlm_lock_get(lock); list_add_tail(&lock->list, &res->blocked); kick_thread = 1; } } - /* reduce the inflight count, this may result in the lockres - * being purged below during calc_usage */ - if (lock->ml.node == dlm->node_num) - dlm_lockres_drop_inflight_ref(dlm, res); spin_unlock(&res->spinlock); wake_up(&res->wq); @@ -231,10 +223,16 @@ static enum dlm_status dlmlock_remote(struct dlm_ctxt *dlm, lock->ml.type, res->lockname.len, res->lockname.name, flags); + /* + * Wait if resource is getting recovered, remastered, etc. + * If the resource was remastered and new owner is self, then exit. + */ spin_lock(&res->spinlock); - - /* will exit this call with spinlock held */ __dlm_wait_on_lockres(res); + if (res->owner == dlm->node_num) { + spin_unlock(&res->spinlock); + return DLM_RECOVERING; + } res->state |= DLM_LOCK_RES_IN_PROGRESS; /* add lock to local (secondary) queue */ @@ -319,27 +317,23 @@ static enum dlm_status dlm_send_remote_lock_request(struct dlm_ctxt *dlm, tmpret = o2net_send_message(DLM_CREATE_LOCK_MSG, dlm->key, &create, sizeof(create), res->owner, &status); if (tmpret >= 0) { - // successfully sent and received - ret = status; // this is already a dlm_status + ret = status; if (ret == DLM_REJECTED) { - mlog(ML_ERROR, "%s:%.*s: BUG. this is a stale lockres " - "no longer owned by %u. that node is coming back " - "up currently.\n", dlm->name, create.namelen, + mlog(ML_ERROR, "%s: res %.*s, Stale lockres no longer " + "owned by node %u. That node is coming back up " + "currently.\n", dlm->name, create.namelen, create.name, res->owner); dlm_print_one_lock_resource(res); BUG(); } } else { - mlog(ML_ERROR, "Error %d when sending message %u (key 0x%x) to " - "node %u\n", tmpret, DLM_CREATE_LOCK_MSG, dlm->key, - res->owner); - if (dlm_is_host_down(tmpret)) { + mlog(ML_ERROR, "%s: res %.*s, Error %d send CREATE LOCK to " + "node %u\n", dlm->name, create.namelen, create.name, + tmpret, res->owner); + if (dlm_is_host_down(tmpret)) ret = DLM_RECOVERING; - mlog(0, "node %u died so returning DLM_RECOVERING " - "from lock message!\n", res->owner); - } else { + else ret = dlm_err_to_dlm_status(tmpret); - } } return ret; @@ -440,7 +434,7 @@ struct dlm_lock * dlm_new_lock(int type, u8 node, u64 cookie, /* zero memory only if kernel-allocated */ lksb = kzalloc(sizeof(*lksb), GFP_NOFS); if (!lksb) { - kfree(lock); + kmem_cache_free(dlm_lock_cache, lock); return NULL; } kernel_allocated = 1; @@ -718,18 +712,10 @@ retry_lock: if (status == DLM_RECOVERING || status == DLM_MIGRATING || status == DLM_FORWARD) { - mlog(0, "retrying lock with migration/" - "recovery/in progress\n"); msleep(100); - /* no waiting for dlm_reco_thread */ if (recovery) { if (status != DLM_RECOVERING) goto retry_lock; - - mlog(0, "%s: got RECOVERING " - "for $RECOVERY lock, master " - "was %u\n", dlm->name, - res->owner); /* wait to see the node go down, then * drop down and allow the lockres to * get cleaned up. need to remaster. */ @@ -741,6 +727,14 @@ retry_lock: } } + /* Inflight taken in dlm_get_lock_resource() is dropped here */ + spin_lock(&res->spinlock); + dlm_lockres_drop_inflight_ref(dlm, res); + spin_unlock(&res->spinlock); + + dlm_lockres_calc_usage(dlm, res); + dlm_kick_thread(dlm, res); + if (status != DLM_NORMAL) { lock->lksb->flags &= ~DLM_LKSB_GET_LVB; if (status != DLM_NOTQUEUED) diff --git a/fs/ocfs2/dlm/dlmmaster.c b/fs/ocfs2/dlm/dlmmaster.c index fede57ed005..82abf0cc9a1 100644 --- a/fs/ocfs2/dlm/dlmmaster.c +++ b/fs/ocfs2/dlm/dlmmaster.c @@ -82,9 +82,9 @@ static inline int dlm_mle_equal(struct dlm_ctxt *dlm, return 1; } -static struct kmem_cache *dlm_lockres_cache = NULL; -static struct kmem_cache *dlm_lockname_cache = NULL; -static struct kmem_cache *dlm_mle_cache = NULL; +static struct kmem_cache *dlm_lockres_cache; +static struct kmem_cache *dlm_lockname_cache; +static struct kmem_cache *dlm_mle_cache; static void dlm_mle_release(struct kref *kref); static void dlm_init_mle(struct dlm_master_list_entry *mle, @@ -342,16 +342,13 @@ static int dlm_find_mle(struct dlm_ctxt *dlm, { struct dlm_master_list_entry *tmpmle; struct hlist_head *bucket; - struct hlist_node *list; unsigned int hash; assert_spin_locked(&dlm->master_lock); hash = dlm_lockid_hash(name, namelen); bucket = dlm_master_hash(dlm, hash); - hlist_for_each(list, bucket) { - tmpmle = hlist_entry(list, struct dlm_master_list_entry, - master_hash_node); + hlist_for_each_entry(tmpmle, bucket, master_hash_node) { if (!dlm_mle_equal(dlm, tmpmle, name, namelen)) continue; dlm_get_mle(tmpmle); @@ -475,11 +472,15 @@ bail: void dlm_destroy_master_caches(void) { - if (dlm_lockname_cache) + if (dlm_lockname_cache) { kmem_cache_destroy(dlm_lockname_cache); + dlm_lockname_cache = NULL; + } - if (dlm_lockres_cache) + if (dlm_lockres_cache) { kmem_cache_destroy(dlm_lockres_cache); + dlm_lockres_cache = NULL; + } } static void dlm_lockres_release(struct kref *kref) @@ -580,6 +581,7 @@ static void dlm_init_lockres(struct dlm_ctxt *dlm, atomic_set(&res->asts_reserved, 0); res->migration_pending = 0; res->inflight_locks = 0; + res->inflight_assert_workers = 0; res->dlm = dlm; @@ -631,42 +633,94 @@ error: return NULL; } -void __dlm_lockres_grab_inflight_ref(struct dlm_ctxt *dlm, - struct dlm_lock_resource *res, - int new_lockres, - const char *file, - int line) +void dlm_lockres_set_refmap_bit(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res, int bit) { - if (!new_lockres) - assert_spin_locked(&res->spinlock); + assert_spin_locked(&res->spinlock); + + mlog(0, "res %.*s, set node %u, %ps()\n", res->lockname.len, + res->lockname.name, bit, __builtin_return_address(0)); + + set_bit(bit, res->refmap); +} + +void dlm_lockres_clear_refmap_bit(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res, int bit) +{ + assert_spin_locked(&res->spinlock); + + mlog(0, "res %.*s, clr node %u, %ps()\n", res->lockname.len, + res->lockname.name, bit, __builtin_return_address(0)); + + clear_bit(bit, res->refmap); +} + + +void dlm_lockres_grab_inflight_ref(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res) +{ + assert_spin_locked(&res->spinlock); - if (!test_bit(dlm->node_num, res->refmap)) { - BUG_ON(res->inflight_locks != 0); - dlm_lockres_set_refmap_bit(dlm->node_num, res); - } res->inflight_locks++; - mlog(0, "%s:%.*s: inflight++: now %u\n", - dlm->name, res->lockname.len, res->lockname.name, - res->inflight_locks); + + mlog(0, "%s: res %.*s, inflight++: now %u, %ps()\n", dlm->name, + res->lockname.len, res->lockname.name, res->inflight_locks, + __builtin_return_address(0)); } -void __dlm_lockres_drop_inflight_ref(struct dlm_ctxt *dlm, - struct dlm_lock_resource *res, - const char *file, - int line) +void dlm_lockres_drop_inflight_ref(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res) { assert_spin_locked(&res->spinlock); BUG_ON(res->inflight_locks == 0); + res->inflight_locks--; - mlog(0, "%s:%.*s: inflight--: now %u\n", - dlm->name, res->lockname.len, res->lockname.name, - res->inflight_locks); - if (res->inflight_locks == 0) - dlm_lockres_clear_refmap_bit(dlm->node_num, res); + + mlog(0, "%s: res %.*s, inflight--: now %u, %ps()\n", dlm->name, + res->lockname.len, res->lockname.name, res->inflight_locks, + __builtin_return_address(0)); + wake_up(&res->wq); } +void __dlm_lockres_grab_inflight_worker(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res) +{ + assert_spin_locked(&res->spinlock); + res->inflight_assert_workers++; + mlog(0, "%s:%.*s: inflight assert worker++: now %u\n", + dlm->name, res->lockname.len, res->lockname.name, + res->inflight_assert_workers); +} + +static void dlm_lockres_grab_inflight_worker(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res) +{ + spin_lock(&res->spinlock); + __dlm_lockres_grab_inflight_worker(dlm, res); + spin_unlock(&res->spinlock); +} + +static void __dlm_lockres_drop_inflight_worker(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res) +{ + assert_spin_locked(&res->spinlock); + BUG_ON(res->inflight_assert_workers == 0); + res->inflight_assert_workers--; + mlog(0, "%s:%.*s: inflight assert worker--: now %u\n", + dlm->name, res->lockname.len, res->lockname.name, + res->inflight_assert_workers); +} + +static void dlm_lockres_drop_inflight_worker(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res) +{ + spin_lock(&res->spinlock); + __dlm_lockres_drop_inflight_worker(dlm, res); + spin_unlock(&res->spinlock); +} + /* * lookup a lock resource by name. * may already exist in the hashtable. @@ -697,7 +751,6 @@ struct dlm_lock_resource * dlm_get_lock_resource(struct dlm_ctxt *dlm, unsigned int hash; int tries = 0; int bit, wait_on_recovery = 0; - int drop_inflight_if_nonlocal = 0; BUG_ON(!lockid); @@ -709,36 +762,33 @@ lookup: spin_lock(&dlm->spinlock); tmpres = __dlm_lookup_lockres_full(dlm, lockid, namelen, hash); if (tmpres) { - int dropping_ref = 0; - spin_unlock(&dlm->spinlock); - spin_lock(&tmpres->spinlock); - /* We wait for the other thread that is mastering the resource */ + /* Wait on the thread that is mastering the resource */ if (tmpres->owner == DLM_LOCK_RES_OWNER_UNKNOWN) { __dlm_wait_on_lockres(tmpres); BUG_ON(tmpres->owner == DLM_LOCK_RES_OWNER_UNKNOWN); + spin_unlock(&tmpres->spinlock); + dlm_lockres_put(tmpres); + tmpres = NULL; + goto lookup; } - if (tmpres->owner == dlm->node_num) { - BUG_ON(tmpres->state & DLM_LOCK_RES_DROPPING_REF); - dlm_lockres_grab_inflight_ref(dlm, tmpres); - } else if (tmpres->state & DLM_LOCK_RES_DROPPING_REF) - dropping_ref = 1; - spin_unlock(&tmpres->spinlock); - - /* wait until done messaging the master, drop our ref to allow - * the lockres to be purged, start over. */ - if (dropping_ref) { - spin_lock(&tmpres->spinlock); - __dlm_wait_on_lockres_flags(tmpres, DLM_LOCK_RES_DROPPING_REF); + /* Wait on the resource purge to complete before continuing */ + if (tmpres->state & DLM_LOCK_RES_DROPPING_REF) { + BUG_ON(tmpres->owner == dlm->node_num); + __dlm_wait_on_lockres_flags(tmpres, + DLM_LOCK_RES_DROPPING_REF); spin_unlock(&tmpres->spinlock); dlm_lockres_put(tmpres); tmpres = NULL; goto lookup; } - mlog(0, "found in hash!\n"); + /* Grab inflight ref to pin the resource */ + dlm_lockres_grab_inflight_ref(dlm, tmpres); + + spin_unlock(&tmpres->spinlock); if (res) dlm_lockres_put(res); res = tmpres; @@ -829,8 +879,8 @@ lookup: * but they might own this lockres. wait on them. */ bit = find_next_bit(dlm->recovery_map, O2NM_MAX_NODES, 0); if (bit < O2NM_MAX_NODES) { - mlog(ML_NOTICE, "%s:%.*s: at least one node (%d) to " - "recover before lock mastery can begin\n", + mlog(0, "%s: res %.*s, At least one node (%d) " + "to recover before lock mastery can begin\n", dlm->name, namelen, (char *)lockid, bit); wait_on_recovery = 1; } @@ -843,12 +893,11 @@ lookup: /* finally add the lockres to its hash bucket */ __dlm_insert_lockres(dlm, res); - /* since this lockres is new it doesn't not require the spinlock */ - dlm_lockres_grab_inflight_ref_new(dlm, res); - /* if this node does not become the master make sure to drop - * this inflight reference below */ - drop_inflight_if_nonlocal = 1; + /* Grab inflight ref to pin the resource */ + spin_lock(&res->spinlock); + dlm_lockres_grab_inflight_ref(dlm, res); + spin_unlock(&res->spinlock); /* get an extra ref on the mle in case this is a BLOCK * if so, the creator of the BLOCK may try to put the last @@ -864,8 +913,8 @@ redo_request: * dlm spinlock would be detectable be a change on the mle, * so we only need to clear out the recovery map once. */ if (dlm_is_recovery_lock(lockid, namelen)) { - mlog(ML_NOTICE, "%s: recovery map is not empty, but " - "must master $RECOVERY lock now\n", dlm->name); + mlog(0, "%s: Recovery map is not empty, but must " + "master $RECOVERY lock now\n", dlm->name); if (!dlm_pre_master_reco_lockres(dlm, res)) wait_on_recovery = 0; else { @@ -883,8 +932,8 @@ redo_request: spin_lock(&dlm->spinlock); bit = find_next_bit(dlm->recovery_map, O2NM_MAX_NODES, 0); if (bit < O2NM_MAX_NODES) { - mlog(ML_NOTICE, "%s:%.*s: at least one node (%d) to " - "recover before lock mastery can begin\n", + mlog(0, "%s: res %.*s, At least one node (%d) " + "to recover before lock mastery can begin\n", dlm->name, namelen, (char *)lockid, bit); wait_on_recovery = 1; } else @@ -913,8 +962,8 @@ redo_request: * yet, keep going until it does. this is how the * master will know that asserts are needed back to * the lower nodes. */ - mlog(0, "%s:%.*s: requests only up to %u but master " - "is %u, keep going\n", dlm->name, namelen, + mlog(0, "%s: res %.*s, Requests only up to %u but " + "master is %u, keep going\n", dlm->name, namelen, lockid, nodenum, mle->master); } } @@ -924,13 +973,12 @@ wait: ret = dlm_wait_for_lock_mastery(dlm, res, mle, &blocked); if (ret < 0) { wait_on_recovery = 1; - mlog(0, "%s:%.*s: node map changed, redo the " - "master request now, blocked=%d\n", - dlm->name, res->lockname.len, + mlog(0, "%s: res %.*s, Node map changed, redo the master " + "request now, blocked=%d\n", dlm->name, res->lockname.len, res->lockname.name, blocked); if (++tries > 20) { - mlog(ML_ERROR, "%s:%.*s: spinning on " - "dlm_wait_for_lock_mastery, blocked=%d\n", + mlog(ML_ERROR, "%s: res %.*s, Spinning on " + "dlm_wait_for_lock_mastery, blocked = %d\n", dlm->name, res->lockname.len, res->lockname.name, blocked); dlm_print_one_lock_resource(res); @@ -940,7 +988,8 @@ wait: goto redo_request; } - mlog(0, "lockres mastered by %u\n", res->owner); + mlog(0, "%s: res %.*s, Mastered by %u\n", dlm->name, res->lockname.len, + res->lockname.name, res->owner); /* make sure we never continue without this */ BUG_ON(res->owner == O2NM_MAX_NODES); @@ -952,8 +1001,6 @@ wait: wake_waiters: spin_lock(&res->spinlock); - if (res->owner != dlm->node_num && drop_inflight_if_nonlocal) - dlm_lockres_drop_inflight_ref(dlm, res); res->state &= ~DLM_LOCK_RES_IN_PROGRESS; spin_unlock(&res->spinlock); wake_up(&res->wq); @@ -1426,9 +1473,7 @@ way_up_top: } if (res->owner == dlm->node_num) { - mlog(0, "%s:%.*s: setting bit %u in refmap\n", - dlm->name, namelen, name, request->node_idx); - dlm_lockres_set_refmap_bit(request->node_idx, res); + dlm_lockres_set_refmap_bit(dlm, res, request->node_idx); spin_unlock(&res->spinlock); response = DLM_MASTER_RESP_YES; if (mle) @@ -1493,10 +1538,8 @@ way_up_top: * go back and clean the mles on any * other nodes */ dispatch_assert = 1; - dlm_lockres_set_refmap_bit(request->node_idx, res); - mlog(0, "%s:%.*s: setting bit %u in refmap\n", - dlm->name, namelen, name, - request->node_idx); + dlm_lockres_set_refmap_bit(dlm, res, + request->node_idx); } else response = DLM_MASTER_RESP_NO; } else { @@ -1598,7 +1641,8 @@ send_response: mlog(ML_ERROR, "failed to dispatch assert master work\n"); response = DLM_MASTER_RESP_ERROR; dlm_lockres_put(res); - } + } else + dlm_lockres_grab_inflight_worker(dlm, res); } else { if (res) dlm_lockres_put(res); @@ -1702,7 +1746,7 @@ again: "lockres, set the bit in the refmap\n", namelen, lockname, to); spin_lock(&res->spinlock); - dlm_lockres_set_refmap_bit(to, res); + dlm_lockres_set_refmap_bit(dlm, res, to); spin_unlock(&res->spinlock); } } @@ -1884,8 +1928,10 @@ ok: * up nodes that this node contacted */ while ((nn = find_next_bit (mle->response_map, O2NM_MAX_NODES, nn+1)) < O2NM_MAX_NODES) { - if (nn != dlm->node_num && nn != assert->node_idx) + if (nn != dlm->node_num && nn != assert->node_idx) { master_request = 1; + break; + } } } mle->master = assert->node_idx; @@ -2016,7 +2062,7 @@ int dlm_dispatch_assert_master(struct dlm_ctxt *dlm, int ignore_higher, u8 request_from, u32 flags) { struct dlm_work_item *item; - item = kzalloc(sizeof(*item), GFP_NOFS); + item = kzalloc(sizeof(*item), GFP_ATOMIC); if (!item) return -ENOMEM; @@ -2111,6 +2157,8 @@ static void dlm_assert_master_worker(struct dlm_work_item *item, void *data) dlm_lockres_release_ast(dlm, res); put: + dlm_lockres_drop_inflight_worker(dlm, res); + dlm_lockres_put(res); mlog(0, "finished with dlm_assert_master_worker\n"); @@ -2187,8 +2235,6 @@ int dlm_drop_lockres_ref(struct dlm_ctxt *dlm, struct dlm_lock_resource *res) namelen = res->lockname.len; BUG_ON(namelen > O2NM_MAX_NAME_LEN); - mlog(0, "%s:%.*s: sending deref to %d\n", - dlm->name, namelen, lockname, res->owner); memset(&deref, 0, sizeof(deref)); deref.node_idx = dlm->node_num; deref.namelen = namelen; @@ -2197,14 +2243,12 @@ int dlm_drop_lockres_ref(struct dlm_ctxt *dlm, struct dlm_lock_resource *res) ret = o2net_send_message(DLM_DEREF_LOCKRES_MSG, dlm->key, &deref, sizeof(deref), res->owner, &r); if (ret < 0) - mlog(ML_ERROR, "Error %d when sending message %u (key 0x%x) to " - "node %u\n", ret, DLM_DEREF_LOCKRES_MSG, dlm->key, - res->owner); + mlog(ML_ERROR, "%s: res %.*s, error %d send DEREF to node %u\n", + dlm->name, namelen, lockname, ret, res->owner); else if (r < 0) { /* BAD. other node says I did not have a ref. */ - mlog(ML_ERROR,"while dropping ref on %s:%.*s " - "(master=%u) got %d.\n", dlm->name, namelen, - lockname, res->owner, r); + 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(); } @@ -2260,7 +2304,7 @@ int dlm_deref_lockres_handler(struct o2net_msg *msg, u32 len, void *data, else { BUG_ON(res->state & DLM_LOCK_RES_DROPPING_REF); if (test_bit(node, res->refmap)) { - dlm_lockres_clear_refmap_bit(node, res); + dlm_lockres_clear_refmap_bit(dlm, res, node); cleared = 1; } } @@ -2320,7 +2364,7 @@ static void dlm_deref_lockres_worker(struct dlm_work_item *item, void *data) BUG_ON(res->state & DLM_LOCK_RES_DROPPING_REF); if (test_bit(node, res->refmap)) { __dlm_wait_on_lockres_flags(res, DLM_LOCK_RES_SETREF_INPROG); - dlm_lockres_clear_refmap_bit(node, res); + dlm_lockres_clear_refmap_bit(dlm, res, node); cleared = 1; } spin_unlock(&res->spinlock); @@ -2339,65 +2383,59 @@ static void dlm_deref_lockres_worker(struct dlm_work_item *item, void *data) dlm_lockres_put(res); } -/* Checks whether the lockres can be migrated. Returns 0 if yes, < 0 - * if not. If 0, numlocks is set to the number of locks in the lockres. +/* + * A migrateable resource is one that is : + * 1. locally mastered, and, + * 2. zero local locks, and, + * 3. one or more non-local locks, or, one or more references + * Returns 1 if yes, 0 if not. */ static int dlm_is_lockres_migrateable(struct dlm_ctxt *dlm, - struct dlm_lock_resource *res, - int *numlocks, - int *hasrefs) + struct dlm_lock_resource *res) { - int ret; - int i; - int count = 0; + enum dlm_lockres_list idx; + int nonlocal = 0, node_ref; struct list_head *queue; struct dlm_lock *lock; + u64 cookie; assert_spin_locked(&res->spinlock); - *numlocks = 0; - *hasrefs = 0; - - ret = -EINVAL; - if (res->owner == DLM_LOCK_RES_OWNER_UNKNOWN) { - mlog(0, "cannot migrate lockres with unknown owner!\n"); - goto leave; - } + /* delay migration when the lockres is in MIGRATING state */ + if (res->state & DLM_LOCK_RES_MIGRATING) + return 0; - if (res->owner != dlm->node_num) { - mlog(0, "cannot migrate lockres this node doesn't own!\n"); - goto leave; - } + if (res->owner != dlm->node_num) + return 0; - ret = 0; - queue = &res->granted; - for (i = 0; i < 3; i++) { + for (idx = DLM_GRANTED_LIST; idx <= DLM_BLOCKED_LIST; idx++) { + queue = dlm_list_idx_to_ptr(res, idx); list_for_each_entry(lock, queue, list) { - ++count; - if (lock->ml.node == dlm->node_num) { - mlog(0, "found a lock owned by this node still " - "on the %s queue! will not migrate this " - "lockres\n", (i == 0 ? "granted" : - (i == 1 ? "converting" : - "blocked"))); - ret = -ENOTEMPTY; - goto leave; + if (lock->ml.node != dlm->node_num) { + nonlocal++; + continue; } + cookie = be64_to_cpu(lock->ml.cookie); + mlog(0, "%s: Not migrateable res %.*s, lock %u:%llu on " + "%s list\n", dlm->name, res->lockname.len, + res->lockname.name, + dlm_get_lock_cookie_node(cookie), + dlm_get_lock_cookie_seq(cookie), + dlm_list_in_text(idx)); + return 0; } - queue++; } - *numlocks = count; - - count = find_next_bit(res->refmap, O2NM_MAX_NODES, 0); - if (count < O2NM_MAX_NODES) - *hasrefs = 1; + if (!nonlocal) { + node_ref = find_next_bit(res->refmap, O2NM_MAX_NODES, 0); + if (node_ref >= O2NM_MAX_NODES) + return 0; + } - mlog(0, "%s: res %.*s, Migrateable, locks %d, refs %d\n", dlm->name, - res->lockname.len, res->lockname.name, *numlocks, *hasrefs); + mlog(0, "%s: res %.*s, Migrateable\n", dlm->name, res->lockname.len, + res->lockname.name); -leave: - return ret; + return 1; } /* @@ -2406,8 +2444,7 @@ leave: static int dlm_migrate_lockres(struct dlm_ctxt *dlm, - struct dlm_lock_resource *res, - u8 target) + struct dlm_lock_resource *res, u8 target) { struct dlm_master_list_entry *mle = NULL; struct dlm_master_list_entry *oldmle = NULL; @@ -2416,37 +2453,20 @@ static int dlm_migrate_lockres(struct dlm_ctxt *dlm, const char *name; unsigned int namelen; int mle_added = 0; - int numlocks, hasrefs; int wake = 0; if (!dlm_grab(dlm)) return -EINVAL; + BUG_ON(target == O2NM_MAX_NODES); + name = res->lockname.name; namelen = res->lockname.len; - mlog(0, "%s: Migrating %.*s to %u\n", dlm->name, namelen, name, target); - - /* - * ensure this lockres is a proper candidate for migration - */ - spin_lock(&res->spinlock); - ret = dlm_is_lockres_migrateable(dlm, res, &numlocks, &hasrefs); - if (ret < 0) { - spin_unlock(&res->spinlock); - goto leave; - } - spin_unlock(&res->spinlock); - - /* no work to do */ - if (numlocks == 0 && !hasrefs) - goto leave; - - /* - * preallocate up front - * if this fails, abort - */ + mlog(0, "%s: Migrating %.*s to node %u\n", dlm->name, namelen, name, + target); + /* preallocate up front. if this fails, abort */ ret = -ENOMEM; mres = (struct dlm_migratable_lockres *) __get_free_page(GFP_NOFS); if (!mres) { @@ -2462,35 +2482,10 @@ static int dlm_migrate_lockres(struct dlm_ctxt *dlm, ret = 0; /* - * find a node to migrate the lockres to - */ - - spin_lock(&dlm->spinlock); - /* pick a new node */ - if (!test_bit(target, dlm->domain_map) || - target >= O2NM_MAX_NODES) { - target = dlm_pick_migration_target(dlm, res); - } - mlog(0, "%s: res %.*s, Node %u chosen for migration\n", dlm->name, - namelen, name, target); - - if (target >= O2NM_MAX_NODES || - !test_bit(target, dlm->domain_map)) { - /* target chosen is not alive */ - ret = -EINVAL; - } - - if (ret) { - spin_unlock(&dlm->spinlock); - goto fail; - } - - mlog(0, "continuing with target = %u\n", target); - - /* * clear any existing master requests and * add the migration mle to the list */ + spin_lock(&dlm->spinlock); spin_lock(&dlm->master_lock); ret = dlm_add_migration_mle(dlm, res, mle, &oldmle, name, namelen, target, dlm->node_num); @@ -2531,6 +2526,7 @@ fail: dlm_put_mle(mle); } else if (mle) { kmem_cache_free(dlm_mle_cache, mle); + mle = NULL; } goto leave; } @@ -2574,6 +2570,9 @@ fail: res->state &= ~DLM_LOCK_RES_MIGRATING; wake = 1; spin_unlock(&res->spinlock); + if (dlm_is_host_down(ret)) + dlm_wait_for_node_death(dlm, target, + DLM_NODE_DEATH_WAIT_MAX); goto leave; } @@ -2649,69 +2648,52 @@ leave: if (wake) wake_up(&res->wq); - /* TODO: cleanup */ if (mres) free_page((unsigned long)mres); dlm_put(dlm); - mlog(0, "returning %d\n", ret); + mlog(0, "%s: Migrating %.*s to %u, returns %d\n", dlm->name, namelen, + name, target, ret); return ret; } #define DLM_MIGRATION_RETRY_MS 100 -/* Should be called only after beginning the domain leave process. +/* + * Should be called only after beginning the domain leave process. * There should not be any remaining locks on nonlocal lock resources, * and there should be no local locks left on locally mastered resources. * * Called with the dlm spinlock held, may drop it to do migration, but * will re-acquire before exit. * - * Returns: 1 if dlm->spinlock was dropped/retaken, 0 if never dropped */ + * Returns: 1 if dlm->spinlock was dropped/retaken, 0 if never dropped + */ int dlm_empty_lockres(struct dlm_ctxt *dlm, struct dlm_lock_resource *res) { int ret; int lock_dropped = 0; - int numlocks, hasrefs; + u8 target = O2NM_MAX_NODES; + + assert_spin_locked(&dlm->spinlock); spin_lock(&res->spinlock); - if (res->owner != dlm->node_num) { - if (!__dlm_lockres_unused(res)) { - mlog(ML_ERROR, "%s:%.*s: this node is not master, " - "trying to free this but locks remain\n", - dlm->name, res->lockname.len, res->lockname.name); - } - spin_unlock(&res->spinlock); - goto leave; - } + if (dlm_is_lockres_migrateable(dlm, res)) + target = dlm_pick_migration_target(dlm, res); + spin_unlock(&res->spinlock); - /* No need to migrate a lockres having no locks */ - ret = dlm_is_lockres_migrateable(dlm, res, &numlocks, &hasrefs); - if (ret >= 0 && numlocks == 0 && !hasrefs) { - spin_unlock(&res->spinlock); + if (target == O2NM_MAX_NODES) goto leave; - } - spin_unlock(&res->spinlock); /* Wheee! Migrate lockres here! Will sleep so drop spinlock. */ spin_unlock(&dlm->spinlock); lock_dropped = 1; - while (1) { - ret = dlm_migrate_lockres(dlm, res, O2NM_MAX_NODES); - if (ret >= 0) - break; - if (ret == -ENOTEMPTY) { - mlog(ML_ERROR, "lockres %.*s still has local locks!\n", - res->lockname.len, res->lockname.name); - BUG(); - } - - mlog(0, "lockres %.*s: migrate failed, " - "retrying\n", res->lockname.len, - res->lockname.name); - msleep(DLM_MIGRATION_RETRY_MS); - } + ret = dlm_migrate_lockres(dlm, res, target); + if (ret) + mlog(0, "%s: res %.*s, Migrate to node %u failed with %d\n", + dlm->name, res->lockname.len, res->lockname.name, + target, ret); spin_lock(&dlm->spinlock); leave: return lock_dropped; @@ -2868,7 +2850,8 @@ static void dlm_remove_nonlocal_locks(struct dlm_ctxt *dlm, BUG_ON(!list_empty(&lock->bast_list)); BUG_ON(lock->ast_pending); BUG_ON(lock->bast_pending); - dlm_lockres_clear_refmap_bit(lock->ml.node, res); + dlm_lockres_clear_refmap_bit(dlm, res, + lock->ml.node); list_del_init(&lock->list); dlm_lock_put(lock); /* In a normal unlock, we would have added a @@ -2889,67 +2872,61 @@ static void dlm_remove_nonlocal_locks(struct dlm_ctxt *dlm, mlog(0, "%s:%.*s: node %u had a ref to this " "migrating lockres, clearing\n", dlm->name, res->lockname.len, res->lockname.name, bit); - dlm_lockres_clear_refmap_bit(bit, res); + dlm_lockres_clear_refmap_bit(dlm, res, bit); } bit++; } } -/* for now this is not too intelligent. we will - * need stats to make this do the right thing. - * this just finds the first lock on one of the - * queues and uses that node as the target. */ +/* + * Pick a node to migrate the lock resource to. This function selects a + * potential target based first on the locks and then on refmap. It skips + * nodes that are in the process of exiting the domain. + */ static u8 dlm_pick_migration_target(struct dlm_ctxt *dlm, struct dlm_lock_resource *res) { - int i; + enum dlm_lockres_list idx; struct list_head *queue = &res->granted; struct dlm_lock *lock; - int nodenum; + int noderef; + u8 nodenum = O2NM_MAX_NODES; assert_spin_locked(&dlm->spinlock); + assert_spin_locked(&res->spinlock); - spin_lock(&res->spinlock); - for (i=0; i<3; i++) { + /* Go through all the locks */ + for (idx = DLM_GRANTED_LIST; idx <= DLM_BLOCKED_LIST; idx++) { + queue = dlm_list_idx_to_ptr(res, idx); list_for_each_entry(lock, queue, list) { - /* up to the caller to make sure this node - * is alive */ - if (lock->ml.node != dlm->node_num) { - spin_unlock(&res->spinlock); - return lock->ml.node; - } + if (lock->ml.node == dlm->node_num) + continue; + if (test_bit(lock->ml.node, dlm->exit_domain_map)) + continue; + nodenum = lock->ml.node; + goto bail; } - queue++; } - nodenum = find_next_bit(res->refmap, O2NM_MAX_NODES, 0); - if (nodenum < O2NM_MAX_NODES) { - spin_unlock(&res->spinlock); - return nodenum; - } - spin_unlock(&res->spinlock); - mlog(0, "have not found a suitable target yet! checking domain map\n"); - - /* ok now we're getting desperate. pick anyone alive. */ - nodenum = -1; + /* Go thru the refmap */ + noderef = -1; while (1) { - nodenum = find_next_bit(dlm->domain_map, - O2NM_MAX_NODES, nodenum+1); - mlog(0, "found %d in domain map\n", nodenum); - if (nodenum >= O2NM_MAX_NODES) + noderef = find_next_bit(res->refmap, O2NM_MAX_NODES, + noderef + 1); + if (noderef >= O2NM_MAX_NODES) break; - if (nodenum != dlm->node_num) { - mlog(0, "picking %d\n", nodenum); - return nodenum; - } + if (noderef == dlm->node_num) + continue; + if (test_bit(noderef, dlm->exit_domain_map)) + continue; + nodenum = noderef; + goto bail; } - mlog(0, "giving up. no master to migrate to\n"); - return DLM_LOCK_RES_OWNER_UNKNOWN; +bail: + return nodenum; } - - /* this is called by the new master once all lockres * data has been received */ static int dlm_do_migrate_request(struct dlm_ctxt *dlm, @@ -2988,9 +2965,9 @@ static int dlm_do_migrate_request(struct dlm_ctxt *dlm, &migrate, sizeof(migrate), nodenum, &status); if (ret < 0) { - mlog(ML_ERROR, "Error %d when sending message %u (key " - "0x%x) to node %u\n", ret, DLM_MIGRATE_REQUEST_MSG, - dlm->key, nodenum); + mlog(ML_ERROR, "%s: res %.*s, Error %d send " + "MIGRATE_REQUEST to node %u\n", dlm->name, + migrate.namelen, migrate.name, ret, nodenum); if (!dlm_is_host_down(ret)) { mlog(ML_ERROR, "unhandled error=%d!\n", ret); BUG(); @@ -3009,7 +2986,7 @@ static int dlm_do_migrate_request(struct dlm_ctxt *dlm, dlm->name, res->lockname.len, res->lockname.name, nodenum); spin_lock(&res->spinlock); - dlm_lockres_set_refmap_bit(nodenum, res); + dlm_lockres_set_refmap_bit(dlm, res, nodenum); spin_unlock(&res->spinlock); } } @@ -3152,11 +3129,15 @@ static int dlm_add_migration_mle(struct dlm_ctxt *dlm, /* remove it so that only one mle will be found */ __dlm_unlink_mle(dlm, tmp); __dlm_mle_detach_hb_events(dlm, tmp); - ret = DLM_MIGRATE_RESPONSE_MASTERY_REF; - mlog(0, "%s:%.*s: master=%u, newmaster=%u, " - "telling master to get ref for cleared out mle " - "during migration\n", dlm->name, namelen, name, - master, new_master); + if (tmp->type == DLM_MLE_MASTER) { + ret = DLM_MIGRATE_RESPONSE_MASTERY_REF; + mlog(0, "%s:%.*s: master=%u, newmaster=%u, " + "telling master to get ref " + "for cleared out mle during " + "migration\n", dlm->name, + namelen, name, master, + new_master); + } } spin_unlock(&tmp->spinlock); } @@ -3254,7 +3235,7 @@ void dlm_clean_master_list(struct dlm_ctxt *dlm, u8 dead_node) struct dlm_master_list_entry *mle; struct dlm_lock_resource *res; struct hlist_head *bucket; - struct hlist_node *list; + struct hlist_node *tmp; unsigned int i; mlog(0, "dlm=%s, dead node=%u\n", dlm->name, dead_node); @@ -3265,10 +3246,7 @@ top: spin_lock(&dlm->master_lock); for (i = 0; i < DLM_HASH_BUCKETS; i++) { bucket = dlm_master_hash(dlm, i); - hlist_for_each(list, bucket) { - mle = hlist_entry(list, struct dlm_master_list_entry, - master_hash_node); - + hlist_for_each_entry_safe(mle, tmp, bucket, master_hash_node) { BUG_ON(mle->type != DLM_MLE_BLOCK && mle->type != DLM_MLE_MASTER && mle->type != DLM_MLE_MIGRATION); @@ -3343,7 +3321,7 @@ int dlm_finish_migration(struct dlm_ctxt *dlm, struct dlm_lock_resource *res, * mastery reference here since old_master will briefly have * a reference after the migration completes */ spin_lock(&res->spinlock); - dlm_lockres_set_refmap_bit(old_master, res); + dlm_lockres_set_refmap_bit(dlm, res, old_master); spin_unlock(&res->spinlock); mlog(0, "now time to do a migrate request to other nodes\n"); @@ -3449,7 +3427,7 @@ void dlm_force_free_mles(struct dlm_ctxt *dlm) int i; struct hlist_head *bucket; struct dlm_master_list_entry *mle; - struct hlist_node *tmp, *list; + struct hlist_node *tmp; /* * We notified all other nodes that we are exiting the domain and @@ -3465,9 +3443,7 @@ void dlm_force_free_mles(struct dlm_ctxt *dlm) for (i = 0; i < DLM_HASH_BUCKETS; i++) { bucket = dlm_master_hash(dlm, i); - hlist_for_each_safe(list, tmp, bucket) { - mle = hlist_entry(list, struct dlm_master_list_entry, - master_hash_node); + hlist_for_each_entry_safe(mle, tmp, bucket, master_hash_node) { if (mle->type != DLM_MLE_BLOCK) { mlog(ML_ERROR, "bad mle: %p\n", mle); dlm_print_one_mle(mle); diff --git a/fs/ocfs2/dlm/dlmrecovery.c b/fs/ocfs2/dlm/dlmrecovery.c index f1beb6fc254..45067faf569 100644 --- a/fs/ocfs2/dlm/dlmrecovery.c +++ b/fs/ocfs2/dlm/dlmrecovery.c @@ -55,9 +55,6 @@ static void dlm_do_local_recovery_cleanup(struct dlm_ctxt *dlm, u8 dead_node); static int dlm_recovery_thread(void *data); -void dlm_complete_recovery_thread(struct dlm_ctxt *dlm); -int dlm_launch_recovery_thread(struct dlm_ctxt *dlm); -void dlm_kick_recovery_thread(struct dlm_ctxt *dlm); static int dlm_do_recovery(struct dlm_ctxt *dlm); static int dlm_pick_recovery_master(struct dlm_ctxt *dlm); @@ -362,40 +359,38 @@ static int dlm_is_node_recovered(struct dlm_ctxt *dlm, u8 node) } -int dlm_wait_for_node_death(struct dlm_ctxt *dlm, u8 node, int timeout) +void dlm_wait_for_node_death(struct dlm_ctxt *dlm, u8 node, int timeout) { - if (timeout) { - mlog(ML_NOTICE, "%s: waiting %dms for notification of " - "death of node %u\n", dlm->name, timeout, node); + if (dlm_is_node_dead(dlm, node)) + return; + + printk(KERN_NOTICE "o2dlm: Waiting on the death of node %u in " + "domain %s\n", node, dlm->name); + + if (timeout) wait_event_timeout(dlm->dlm_reco_thread_wq, - dlm_is_node_dead(dlm, node), - msecs_to_jiffies(timeout)); - } else { - mlog(ML_NOTICE, "%s: waiting indefinitely for notification " - "of death of node %u\n", dlm->name, node); + dlm_is_node_dead(dlm, node), + msecs_to_jiffies(timeout)); + else wait_event(dlm->dlm_reco_thread_wq, dlm_is_node_dead(dlm, node)); - } - /* for now, return 0 */ - return 0; } -int dlm_wait_for_node_recovery(struct dlm_ctxt *dlm, u8 node, int timeout) +void dlm_wait_for_node_recovery(struct dlm_ctxt *dlm, u8 node, int timeout) { - if (timeout) { - mlog(0, "%s: waiting %dms for notification of " - "recovery of node %u\n", dlm->name, timeout, node); + if (dlm_is_node_recovered(dlm, node)) + return; + + printk(KERN_NOTICE "o2dlm: Waiting on the recovery of node %u in " + "domain %s\n", node, dlm->name); + + if (timeout) wait_event_timeout(dlm->dlm_reco_thread_wq, - dlm_is_node_recovered(dlm, node), - msecs_to_jiffies(timeout)); - } else { - mlog(0, "%s: waiting indefinitely for notification " - "of recovery of node %u\n", dlm->name, node); + dlm_is_node_recovered(dlm, node), + msecs_to_jiffies(timeout)); + else wait_event(dlm->dlm_reco_thread_wq, dlm_is_node_recovered(dlm, node)); - } - /* for now, return 0 */ - return 0; } /* callers of the top-level api calls (dlmlock/dlmunlock) should @@ -430,6 +425,8 @@ static void dlm_begin_recovery(struct dlm_ctxt *dlm) { spin_lock(&dlm->spinlock); BUG_ON(dlm->reco.state & DLM_RECO_STATE_ACTIVE); + printk(KERN_NOTICE "o2dlm: Begin recovery on domain %s for node %u\n", + dlm->name, dlm->reco.dead_node); dlm->reco.state |= DLM_RECO_STATE_ACTIVE; spin_unlock(&dlm->spinlock); } @@ -440,9 +437,18 @@ static void dlm_end_recovery(struct dlm_ctxt *dlm) BUG_ON(!(dlm->reco.state & DLM_RECO_STATE_ACTIVE)); dlm->reco.state &= ~DLM_RECO_STATE_ACTIVE; spin_unlock(&dlm->spinlock); + printk(KERN_NOTICE "o2dlm: End recovery on domain %s\n", dlm->name); wake_up(&dlm->reco.event); } +static void dlm_print_recovery_master(struct dlm_ctxt *dlm) +{ + printk(KERN_NOTICE "o2dlm: Node %u (%s) is the Recovery Master for the " + "dead node %u in domain %s\n", dlm->reco.new_master, + (dlm->node_num == dlm->reco.new_master ? "me" : "he"), + dlm->reco.dead_node, dlm->name); +} + static int dlm_do_recovery(struct dlm_ctxt *dlm) { int status = 0; @@ -505,9 +511,8 @@ static int dlm_do_recovery(struct dlm_ctxt *dlm) } mlog(0, "another node will master this recovery session.\n"); } - mlog(0, "dlm=%s (%d), new_master=%u, this node=%u, dead_node=%u\n", - dlm->name, task_pid_nr(dlm->dlm_reco_thread_task), dlm->reco.new_master, - dlm->node_num, dlm->reco.dead_node); + + dlm_print_recovery_master(dlm); /* it is safe to start everything back up here * because all of the dead node's lock resources @@ -518,15 +523,13 @@ static int dlm_do_recovery(struct dlm_ctxt *dlm) return 0; master_here: - mlog(ML_NOTICE, "(%d) Node %u is the Recovery Master for the Dead Node " - "%u for Domain %s\n", task_pid_nr(dlm->dlm_reco_thread_task), - dlm->node_num, dlm->reco.dead_node, dlm->name); + dlm_print_recovery_master(dlm); status = dlm_remaster_locks(dlm, dlm->reco.dead_node); if (status < 0) { /* we should never hit this anymore */ - mlog(ML_ERROR, "error %d remastering locks for node %u, " - "retrying.\n", status, dlm->reco.dead_node); + mlog(ML_ERROR, "%s: Error %d remastering locks for node %u, " + "retrying.\n", dlm->name, status, dlm->reco.dead_node); /* yield a bit to allow any final network messages * to get handled on remaining nodes */ msleep(100); @@ -534,7 +537,10 @@ master_here: /* success! see if any other nodes need recovery */ mlog(0, "DONE mastering recovery of %s:%u here(this=%u)!\n", dlm->name, dlm->reco.dead_node, dlm->node_num); - dlm_reset_recovery(dlm); + spin_lock(&dlm->spinlock); + __dlm_reset_recovery(dlm); + dlm->reco.state &= ~DLM_RECO_STATE_FINALIZE; + spin_unlock(&dlm->spinlock); } dlm_end_recovery(dlm); @@ -567,7 +573,7 @@ static int dlm_remaster_locks(struct dlm_ctxt *dlm, u8 dead_node) BUG_ON(ndata->state != DLM_RECO_NODE_DATA_INIT); ndata->state = DLM_RECO_NODE_DATA_REQUESTING; - mlog(0, "requesting lock info from node %u\n", + mlog(0, "%s: Requesting lock info from node %u\n", dlm->name, ndata->node_num); if (ndata->node_num == dlm->node_num) { @@ -640,7 +646,7 @@ static int dlm_remaster_locks(struct dlm_ctxt *dlm, u8 dead_node) spin_unlock(&dlm_reco_state_lock); } - mlog(0, "done requesting all lock info\n"); + mlog(0, "%s: Done requesting all lock info\n", dlm->name); /* nodes should be sending reco data now * just need to wait */ @@ -692,6 +698,14 @@ static int dlm_remaster_locks(struct dlm_ctxt *dlm, u8 dead_node) if (all_nodes_done) { int ret; + /* Set this flag on recovery master to avoid + * a new recovery for another dead node start + * before the recovery is not done. That may + * cause recovery hung.*/ + spin_lock(&dlm->spinlock); + dlm->reco.state |= DLM_RECO_STATE_FINALIZE; + spin_unlock(&dlm->spinlock); + /* all nodes are now in DLM_RECO_NODE_DATA_DONE state * just send a finalize message to everyone and * clean up */ @@ -783,7 +797,8 @@ static int dlm_request_all_locks(struct dlm_ctxt *dlm, u8 request_from, u8 dead_node) { struct dlm_lock_request lr; - enum dlm_status ret; + int ret; + int status; mlog(0, "\n"); @@ -796,16 +811,16 @@ static int dlm_request_all_locks(struct dlm_ctxt *dlm, u8 request_from, lr.dead_node = dead_node; // send message - ret = DLM_NOLOCKMGR; ret = o2net_send_message(DLM_LOCK_REQUEST_MSG, dlm->key, - &lr, sizeof(lr), request_from, NULL); + &lr, sizeof(lr), request_from, &status); /* negative status is handled by caller */ if (ret < 0) - mlog(ML_ERROR, "Error %d when sending message %u (key " - "0x%x) to node %u\n", ret, DLM_LOCK_REQUEST_MSG, - dlm->key, request_from); - + mlog(ML_ERROR, "%s: Error %d send LOCK_REQUEST to node %u " + "to recover dead node %u\n", dlm->name, ret, + request_from, dead_node); + else + ret = status; // return from here, then // sleep until all received or error return ret; @@ -956,9 +971,9 @@ static int dlm_send_all_done_msg(struct dlm_ctxt *dlm, u8 dead_node, u8 send_to) ret = o2net_send_message(DLM_RECO_DATA_DONE_MSG, dlm->key, &done_msg, sizeof(done_msg), send_to, &tmpret); if (ret < 0) { - mlog(ML_ERROR, "Error %d when sending message %u (key " - "0x%x) to node %u\n", ret, DLM_RECO_DATA_DONE_MSG, - dlm->key, send_to); + mlog(ML_ERROR, "%s: Error %d send RECO_DATA_DONE to node %u " + "to recover dead node %u\n", dlm->name, ret, send_to, + dead_node); if (!dlm_is_host_down(ret)) { BUG(); } @@ -1127,9 +1142,11 @@ static int dlm_send_mig_lockres_msg(struct dlm_ctxt *dlm, if (ret < 0) { /* XXX: negative status is not handled. * this will end up killing this node. */ - mlog(ML_ERROR, "Error %d when sending message %u (key " - "0x%x) to node %u\n", ret, DLM_MIG_LOCKRES_MSG, - dlm->key, send_to); + mlog(ML_ERROR, "%s: res %.*s, Error %d send MIG_LOCKRES to " + "node %u (%s)\n", dlm->name, mres->lockname_len, + mres->lockname, ret, send_to, + (orig_flags & DLM_MRES_MIGRATION ? + "migration" : "recovery")); } else { /* might get an -ENOMEM back here */ ret = status; @@ -1401,6 +1418,7 @@ int dlm_mig_lockres_handler(struct o2net_msg *msg, u32 len, void *data, mres->lockname_len, mres->lockname); ret = -EFAULT; spin_unlock(&res->spinlock); + dlm_lockres_put(res); goto leave; } res->state |= DLM_LOCK_RES_MIGRATING; @@ -1491,10 +1509,8 @@ leave: dlm_put(dlm); if (ret < 0) { - if (buf) - kfree(buf); - if (item) - kfree(item); + kfree(buf); + kfree(item); mlog_errno(ret); } @@ -1692,7 +1708,8 @@ int dlm_master_requery_handler(struct o2net_msg *msg, u32 len, void *data, mlog_errno(-ENOMEM); /* retry!? */ BUG(); - } + } else + __dlm_lockres_grab_inflight_worker(dlm, res); } else /* put.. incase we are not the master */ dlm_lockres_put(res); spin_unlock(&res->spinlock); @@ -1745,13 +1762,13 @@ static int dlm_process_recovery_data(struct dlm_ctxt *dlm, struct dlm_migratable_lockres *mres) { struct dlm_migratable_lock *ml; - struct list_head *queue; + struct list_head *queue, *iter; struct list_head *tmpq = NULL; struct dlm_lock *newlock = NULL; struct dlm_lockstatus *lksb = NULL; int ret = 0; int i, j, bad; - struct dlm_lock *lock = NULL; + struct dlm_lock *lock; u8 from = O2NM_MAX_NODES; unsigned int added = 0; __be64 c; @@ -1767,7 +1784,7 @@ static int dlm_process_recovery_data(struct dlm_ctxt *dlm, dlm->name, mres->lockname_len, mres->lockname, from); spin_lock(&res->spinlock); - dlm_lockres_set_refmap_bit(from, res); + dlm_lockres_set_refmap_bit(dlm, res, from); spin_unlock(&res->spinlock); added++; break; @@ -1786,14 +1803,16 @@ static int dlm_process_recovery_data(struct dlm_ctxt *dlm, /* MIGRATION ONLY! */ BUG_ON(!(mres->flags & DLM_MRES_MIGRATION)); + lock = NULL; spin_lock(&res->spinlock); for (j = DLM_GRANTED_LIST; j <= DLM_BLOCKED_LIST; j++) { tmpq = dlm_list_idx_to_ptr(res, j); - list_for_each_entry(lock, tmpq, list) { - if (lock->ml.cookie != ml->cookie) - lock = NULL; - else + list_for_each(iter, tmpq) { + lock = list_entry(iter, + struct dlm_lock, list); + if (lock->ml.cookie == ml->cookie) break; + lock = NULL; } if (lock) break; @@ -1881,6 +1900,13 @@ static int dlm_process_recovery_data(struct dlm_ctxt *dlm, if (ml->type == LKM_NLMODE) goto skip_lvb; + /* + * If the lock is in the blocked list it can't have a valid lvb, + * so skip it + */ + if (ml->list == DLM_BLOCKED_LIST) + goto skip_lvb; + if (!dlm_lvb_is_empty(mres->lvb)) { if (lksb->flags & DLM_LKSB_PUT_LVB) { /* other node was trying to update @@ -1961,11 +1987,19 @@ skip_lvb: } if (!bad) { dlm_lock_get(newlock); - list_add_tail(&newlock->list, queue); + if (mres->flags & DLM_MRES_RECOVERY && + ml->list == DLM_CONVERTING_LIST && + newlock->ml.type > + newlock->ml.convert_type) { + /* newlock is doing downconvert, add it to the + * head of converting list */ + list_add(&newlock->list, queue); + } else + list_add_tail(&newlock->list, queue); mlog(0, "%s:%.*s: added lock for node %u, " "setting refmap bit\n", dlm->name, res->lockname.len, res->lockname.name, ml->node); - dlm_lockres_set_refmap_bit(ml->node, res); + dlm_lockres_set_refmap_bit(dlm, res, ml->node); added++; } spin_unlock(&res->spinlock); @@ -2076,7 +2110,6 @@ static void dlm_finish_local_lockres_recovery(struct dlm_ctxt *dlm, u8 dead_node, u8 new_master) { int i; - struct hlist_node *hash_iter; struct hlist_head *bucket; struct dlm_lock_resource *res, *next; @@ -2084,6 +2117,9 @@ static void dlm_finish_local_lockres_recovery(struct dlm_ctxt *dlm, list_for_each_entry_safe(res, next, &dlm->reco.resources, recovering) { if (res->owner == dead_node) { + mlog(0, "%s: res %.*s, Changing owner from %u to %u\n", + dlm->name, res->lockname.len, res->lockname.name, + res->owner, new_master); list_del_init(&res->recovering); spin_lock(&res->spinlock); /* new_master has our reference from @@ -2104,41 +2140,31 @@ static void dlm_finish_local_lockres_recovery(struct dlm_ctxt *dlm, * if necessary */ for (i = 0; i < DLM_HASH_BUCKETS; i++) { bucket = dlm_lockres_hash(dlm, i); - hlist_for_each_entry(res, hash_iter, bucket, hash_node) { - if (res->state & DLM_LOCK_RES_RECOVERING) { - if (res->owner == dead_node) { - mlog(0, "(this=%u) res %.*s owner=%u " - "was not on recovering list, but " - "clearing state anyway\n", - dlm->node_num, res->lockname.len, - res->lockname.name, new_master); - } else if (res->owner == dlm->node_num) { - mlog(0, "(this=%u) res %.*s owner=%u " - "was not on recovering list, " - "owner is THIS node, clearing\n", - dlm->node_num, res->lockname.len, - res->lockname.name, new_master); - } else - continue; + hlist_for_each_entry(res, bucket, hash_node) { + if (!(res->state & DLM_LOCK_RES_RECOVERING)) + continue; - if (!list_empty(&res->recovering)) { - mlog(0, "%s:%.*s: lockres was " - "marked RECOVERING, owner=%u\n", - dlm->name, res->lockname.len, - res->lockname.name, res->owner); - list_del_init(&res->recovering); - dlm_lockres_put(res); - } - spin_lock(&res->spinlock); - /* new_master has our reference from - * the lock state sent during recovery */ - dlm_change_lockres_owner(dlm, res, new_master); - res->state &= ~DLM_LOCK_RES_RECOVERING; - if (__dlm_lockres_has_locks(res)) - __dlm_dirty_lockres(dlm, res); - spin_unlock(&res->spinlock); - wake_up(&res->wq); + if (res->owner != dead_node && + res->owner != dlm->node_num) + continue; + + if (!list_empty(&res->recovering)) { + list_del_init(&res->recovering); + dlm_lockres_put(res); } + + /* new_master has our reference from + * the lock state sent during recovery */ + mlog(0, "%s: res %.*s, Changing owner from %u to %u\n", + dlm->name, res->lockname.len, res->lockname.name, + res->owner, new_master); + spin_lock(&res->spinlock); + dlm_change_lockres_owner(dlm, res, new_master); + res->state &= ~DLM_LOCK_RES_RECOVERING; + if (__dlm_lockres_has_locks(res)) + __dlm_dirty_lockres(dlm, res); + spin_unlock(&res->spinlock); + wake_up(&res->wq); } } } @@ -2252,12 +2278,12 @@ static void dlm_free_dead_locks(struct dlm_ctxt *dlm, res->lockname.len, res->lockname.name, freed, dead_node); __dlm_print_one_lock_resource(res); } - dlm_lockres_clear_refmap_bit(dead_node, res); + dlm_lockres_clear_refmap_bit(dlm, res, dead_node); } else if (test_bit(dead_node, res->refmap)) { mlog(0, "%s:%.*s: dead node %u had a ref, but had " "no locks and had not purged before dying\n", dlm->name, res->lockname.len, res->lockname.name, dead_node); - dlm_lockres_clear_refmap_bit(dead_node, res); + dlm_lockres_clear_refmap_bit(dlm, res, dead_node); } /* do not kick thread yet */ @@ -2273,7 +2299,6 @@ static void dlm_free_dead_locks(struct dlm_ctxt *dlm, static void dlm_do_local_recovery_cleanup(struct dlm_ctxt *dlm, u8 dead_node) { - struct hlist_node *iter; struct dlm_lock_resource *res; int i; struct hlist_head *bucket; @@ -2299,7 +2324,7 @@ 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, iter, bucket, hash_node) { + hlist_for_each_entry(res, 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, @@ -2324,9 +2349,9 @@ static void dlm_do_local_recovery_cleanup(struct dlm_ctxt *dlm, u8 dead_node) dlm_revalidate_lvb(dlm, res, dead_node); if (res->owner == dead_node) { if (res->state & DLM_LOCK_RES_DROPPING_REF) { - mlog(ML_NOTICE, "Ignore %.*s for " + mlog(ML_NOTICE, "%s: res %.*s, Skip " "recovery as it is being freed\n", - res->lockname.len, + dlm->name, res->lockname.len, res->lockname.name); } else dlm_move_lockres_to_recovery_list(dlm, @@ -2335,6 +2360,14 @@ static void dlm_do_local_recovery_cleanup(struct dlm_ctxt *dlm, u8 dead_node) } else if (res->owner == dlm->node_num) { dlm_free_dead_locks(dlm, res, dead_node); __dlm_lockres_calc_usage(dlm, res); + } else if (res->owner == DLM_LOCK_RES_OWNER_UNKNOWN) { + if (test_bit(dead_node, res->refmap)) { + mlog(0, "%s:%.*s: dead node %u had a ref, but had " + "no locks and had not purged before dying\n", + dlm->name, res->lockname.len, + res->lockname.name, dead_node); + dlm_lockres_clear_refmap_bit(dlm, res, dead_node); + } } spin_unlock(&res->spinlock); } @@ -2393,6 +2426,7 @@ static void __dlm_hb_node_down(struct dlm_ctxt *dlm, int idx) mlog(0, "node %u being removed from domain map!\n", idx); clear_bit(idx, dlm->domain_map); + clear_bit(idx, dlm->exit_domain_map); /* wake up migration waiters if a node goes down. * perhaps later we can genericize this for other waiters. */ wake_up(&dlm->migration_wq); @@ -2698,6 +2732,7 @@ int dlm_begin_reco_handler(struct o2net_msg *msg, u32 len, void *data, dlm->name, br->node_idx, br->dead_node, dlm->reco.dead_node, dlm->reco.new_master); spin_unlock(&dlm->spinlock); + dlm_put(dlm); return -EAGAIN; } spin_unlock(&dlm->spinlock); @@ -2869,8 +2904,8 @@ int dlm_finalize_reco_handler(struct o2net_msg *msg, u32 len, void *data, BUG(); } dlm->reco.state &= ~DLM_RECO_STATE_FINALIZE; + __dlm_reset_recovery(dlm); spin_unlock(&dlm->spinlock); - dlm_reset_recovery(dlm); dlm_kick_recovery_thread(dlm); break; default: diff --git a/fs/ocfs2/dlm/dlmthread.c b/fs/ocfs2/dlm/dlmthread.c index 1d6d1d22c47..69aac6f088a 100644 --- a/fs/ocfs2/dlm/dlmthread.c +++ b/fs/ocfs2/dlm/dlmthread.c @@ -94,24 +94,26 @@ int __dlm_lockres_unused(struct dlm_lock_resource *res) { int bit; + assert_spin_locked(&res->spinlock); + if (__dlm_lockres_has_locks(res)) return 0; + /* Locks are in the process of being created */ + if (res->inflight_locks) + return 0; + if (!list_empty(&res->dirty) || res->state & DLM_LOCK_RES_DIRTY) return 0; if (res->state & DLM_LOCK_RES_RECOVERING) return 0; + /* Another node has this resource with this node as the master */ bit = find_next_bit(res->refmap, O2NM_MAX_NODES, 0); if (bit < O2NM_MAX_NODES) return 0; - /* - * since the bit for dlm->node_num is not set, inflight_locks better - * be zero - */ - BUG_ON(res->inflight_locks != 0); return 1; } @@ -185,8 +187,6 @@ static void dlm_purge_lockres(struct dlm_ctxt *dlm, /* clear our bit from the master's refmap, ignore errors */ ret = dlm_drop_lockres_ref(dlm, res); if (ret < 0) { - mlog(ML_ERROR, "%s: deref %.*s failed %d\n", dlm->name, - res->lockname.len, res->lockname.name, ret); if (!dlm_is_host_down(ret)) BUG(); } @@ -209,7 +209,7 @@ static void dlm_purge_lockres(struct dlm_ctxt *dlm, BUG(); } - __dlm_unhash_lockres(res); + __dlm_unhash_lockres(dlm, res); /* lockres is not in the hash now. drop the flag and wake up * any processes waiting in dlm_get_lock_resource. */ @@ -259,12 +259,15 @@ static void dlm_run_purge_list(struct dlm_ctxt *dlm, * refs on it. */ unused = __dlm_lockres_unused(lockres); if (!unused || - (lockres->state & DLM_LOCK_RES_MIGRATING)) { + (lockres->state & DLM_LOCK_RES_MIGRATING) || + (lockres->inflight_assert_workers != 0)) { mlog(0, "%s: res %.*s is in use or being remastered, " - "used %d, state %d\n", dlm->name, - lockres->lockname.len, lockres->lockname.name, - !unused, lockres->state); - list_move_tail(&dlm->purge_list, &lockres->purge); + "used %d, state %d, assert master workers %u\n", + dlm->name, lockres->lockname.len, + lockres->lockname.name, + !unused, lockres->state, + lockres->inflight_assert_workers); + list_move_tail(&lockres->purge, &dlm->purge_list); spin_unlock(&lockres->spinlock); continue; } @@ -286,8 +289,6 @@ static void dlm_shuffle_lists(struct dlm_ctxt *dlm, struct dlm_lock_resource *res) { struct dlm_lock *lock, *target; - struct list_head *iter; - struct list_head *head; int can_grant = 1; /* @@ -314,9 +315,7 @@ converting: dlm->name, res->lockname.len, res->lockname.name); BUG(); } - head = &res->granted; - list_for_each(iter, head) { - lock = list_entry(iter, struct dlm_lock, list); + list_for_each_entry(lock, &res->granted, list) { if (lock==target) continue; if (!dlm_lock_compatible(lock->ml.type, @@ -333,9 +332,8 @@ converting: target->ml.convert_type; } } - head = &res->converting; - list_for_each(iter, head) { - lock = list_entry(iter, struct dlm_lock, list); + + list_for_each_entry(lock, &res->converting, list) { if (lock==target) continue; if (!dlm_lock_compatible(lock->ml.type, @@ -384,9 +382,7 @@ blocked: goto leave; target = list_entry(res->blocked.next, struct dlm_lock, list); - head = &res->granted; - list_for_each(iter, head) { - lock = list_entry(iter, struct dlm_lock, list); + list_for_each_entry(lock, &res->granted, list) { if (lock==target) continue; if (!dlm_lock_compatible(lock->ml.type, target->ml.type)) { @@ -400,9 +396,7 @@ blocked: } } - head = &res->converting; - list_for_each(iter, head) { - lock = list_entry(iter, struct dlm_lock, list); + list_for_each_entry(lock, &res->converting, list) { if (lock==target) continue; if (!dlm_lock_compatible(lock->ml.type, target->ml.type)) { diff --git a/fs/ocfs2/dlm/dlmunlock.c b/fs/ocfs2/dlm/dlmunlock.c index 850aa7e8753..2e3c9dbab68 100644 --- a/fs/ocfs2/dlm/dlmunlock.c +++ b/fs/ocfs2/dlm/dlmunlock.c @@ -191,7 +191,9 @@ static enum dlm_status dlmunlock_common(struct dlm_ctxt *dlm, DLM_UNLOCK_CLEAR_CONVERT_TYPE); } else if (status == DLM_RECOVERING || status == DLM_MIGRATING || - status == DLM_FORWARD) { + status == DLM_FORWARD || + status == DLM_NOLOCKMGR + ) { /* must clear the actions because this unlock * is about to be retried. cannot free or do * any list manipulation. */ @@ -200,7 +202,8 @@ static enum dlm_status dlmunlock_common(struct dlm_ctxt *dlm, res->lockname.name, status==DLM_RECOVERING?"recovering": (status==DLM_MIGRATING?"migrating": - "forward")); + (status == DLM_FORWARD ? "forward" : + "nolockmanager"))); actions = 0; } if (flags & LKM_CANCEL) @@ -364,7 +367,10 @@ static enum dlm_status dlm_send_remote_unlock_request(struct dlm_ctxt *dlm, * updated state to the recovery master. this thread * just needs to finish out the operation and call * the unlockast. */ - ret = DLM_NORMAL; + if (dlm_is_node_dead(dlm, owner)) + ret = DLM_NORMAL; + else + ret = DLM_NOLOCKMGR; } else { /* something bad. this will BUG in ocfs2 */ ret = dlm_err_to_dlm_status(tmpret); @@ -388,7 +394,6 @@ int dlm_unlock_lock_handler(struct o2net_msg *msg, u32 len, void *data, struct dlm_ctxt *dlm = data; struct dlm_unlock_lock *unlock = (struct dlm_unlock_lock *)msg->buf; struct dlm_lock_resource *res = NULL; - struct list_head *iter; struct dlm_lock *lock = NULL; enum dlm_status status = DLM_NORMAL; int found = 0, i; @@ -458,8 +463,7 @@ int dlm_unlock_lock_handler(struct o2net_msg *msg, u32 len, void *data, } for (i=0; i<3; i++) { - list_for_each(iter, queue) { - lock = list_entry(iter, struct dlm_lock, list); + list_for_each_entry(lock, queue, list) { if (lock->ml.cookie == unlock->cookie && lock->ml.node == unlock->node_idx) { dlm_lock_get(lock); @@ -640,7 +644,9 @@ retry: if (status == DLM_RECOVERING || status == DLM_MIGRATING || - status == DLM_FORWARD) { + status == DLM_FORWARD || + status == DLM_NOLOCKMGR) { + /* We want to go away for a tiny bit to allow recovery * / migration to complete on this resource. I don't * know of any wait queue we could sleep on as this @@ -652,7 +658,7 @@ retry: msleep(50); mlog(0, "retrying unlock due to pending recovery/" - "migration/in-progress\n"); + "migration/in-progress/reconnect\n"); goto retry; } diff --git a/fs/ocfs2/dlm/dlmver.c b/fs/ocfs2/dlm/dlmver.c deleted file mode 100644 index dfc0da4d158..00000000000 --- a/fs/ocfs2/dlm/dlmver.c +++ /dev/null @@ -1,42 +0,0 @@ -/* -*- mode: c; c-basic-offset: 8; -*- - * vim: noexpandtab sw=8 ts=8 sts=0: - * - * dlmver.c - * - * version string - * - * Copyright (C) 2002, 2005 Oracle. All rights reserved. - * - * This program is free software; you can redistribute it and/or - * modify it under the terms of the GNU General Public - * License as published by the Free Software Foundation; either - * version 2 of the License, or (at your option) any later version. - * - * This program is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * General Public License for more details. - * - * You should have received a copy of the GNU General Public - * License along with this program; if not, write to the - * Free Software Foundation, Inc., 59 Temple Place - Suite 330, - * Boston, MA 021110-1307, USA. - */ - -#include <linux/module.h> -#include <linux/kernel.h> - -#include "dlmver.h" - -#define DLM_BUILD_VERSION "1.5.0" - -#define VERSION_STR "OCFS2 DLM " DLM_BUILD_VERSION - -void dlm_print_version(void) -{ - printk(KERN_INFO "%s\n", VERSION_STR); -} - -MODULE_DESCRIPTION(VERSION_STR); - -MODULE_VERSION(DLM_BUILD_VERSION); diff --git a/fs/ocfs2/dlm/dlmver.h b/fs/ocfs2/dlm/dlmver.h deleted file mode 100644 index f674aee77a1..00000000000 --- a/fs/ocfs2/dlm/dlmver.h +++ /dev/null @@ -1,31 +0,0 @@ -/* -*- mode: c; c-basic-offset: 8; -*- - * vim: noexpandtab sw=8 ts=8 sts=0: - * - * dlmfsver.h - * - * Function prototypes - * - * Copyright (C) 2005 Oracle. All rights reserved. - * - * This program is free software; you can redistribute it and/or - * modify it under the terms of the GNU General Public - * License as published by the Free Software Foundation; either - * version 2 of the License, or (at your option) any later version. - * - * This program is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * General Public License for more details. - * - * You should have received a copy of the GNU General Public - * License along with this program; if not, write to the - * Free Software Foundation, Inc., 59 Temple Place - Suite 330, - * Boston, MA 021110-1307, USA. - */ - -#ifndef DLM_VER_H -#define DLM_VER_H - -void dlm_print_version(void); - -#endif /* DLM_VER_H */ diff --git a/fs/ocfs2/dlmfs/Makefile b/fs/ocfs2/dlmfs/Makefile index f14be89a670..eed3db8c5b4 100644 --- a/fs/ocfs2/dlmfs/Makefile +++ b/fs/ocfs2/dlmfs/Makefile @@ -2,4 +2,4 @@ ccflags-y := -Ifs/ocfs2 obj-$(CONFIG_OCFS2_FS) += ocfs2_dlmfs.o -ocfs2_dlmfs-objs := userdlm.o dlmfs.o dlmfsver.o +ocfs2_dlmfs-objs := userdlm.o dlmfs.o diff --git a/fs/ocfs2/dlmfs/dlmfs.c b/fs/ocfs2/dlmfs/dlmfs.c index 8c5c0eddc36..09b7d9dac71 100644 --- a/fs/ocfs2/dlmfs/dlmfs.c +++ b/fs/ocfs2/dlmfs/dlmfs.c @@ -49,7 +49,6 @@ #include "stackglue.h" #include "userdlm.h" -#include "dlmfsver.h" #define MLOG_MASK_PREFIX ML_DLMFS #include "cluster/masklog.h" @@ -88,7 +87,7 @@ struct workqueue_struct *user_dlm_worker; * signifies a bast fired on the lock. */ #define DLMFS_CAPABILITIES "bast stackglue" -extern int param_set_dlmfs_capabilities(const char *val, +static int param_set_dlmfs_capabilities(const char *val, struct kernel_param *kp) { printk(KERN_ERR "%s: readonly parameter\n", kp->name); @@ -224,7 +223,7 @@ static int dlmfs_file_setattr(struct dentry *dentry, struct iattr *attr) static unsigned int dlmfs_file_poll(struct file *file, poll_table *wait) { int event = 0; - struct inode *inode = file->f_path.dentry->d_inode; + struct inode *inode = file_inode(file); struct dlmfs_inode_private *ip = DLMFS_I(inode); poll_wait(file, &ip->ip_lockres.l_event, wait); @@ -245,7 +244,7 @@ static ssize_t dlmfs_file_read(struct file *filp, int bytes_left; ssize_t readlen, got; char *lvb_buf; - struct inode *inode = filp->f_path.dentry->d_inode; + struct inode *inode = file_inode(filp); mlog(0, "inode %lu, count = %zu, *ppos = %llu\n", inode->i_ino, count, *ppos); @@ -293,7 +292,7 @@ static ssize_t dlmfs_file_write(struct file *filp, int bytes_left; ssize_t writelen; char *lvb_buf; - struct inode *inode = filp->f_path.dentry->d_inode; + struct inode *inode = file_inode(filp); mlog(0, "inode %lu, count = %zu, *ppos = %llu\n", inode->i_ino, count, *ppos); @@ -354,7 +353,6 @@ static struct inode *dlmfs_alloc_inode(struct super_block *sb) static void dlmfs_i_callback(struct rcu_head *head) { struct inode *inode = container_of(head, struct inode, i_rcu); - INIT_LIST_HEAD(&inode->i_dentry); kmem_cache_free(dlmfs_inode_cache, DLMFS_I(inode)); } @@ -368,7 +366,7 @@ static void dlmfs_evict_inode(struct inode *inode) int status; struct dlmfs_inode_private *ip; - end_writeback(inode); + clear_inode(inode); mlog(0, "inode %lu\n", inode->i_ino); @@ -401,16 +399,11 @@ static struct backing_dev_info dlmfs_backing_dev_info = { static struct inode *dlmfs_get_root_inode(struct super_block *sb) { struct inode *inode = new_inode(sb); - int mode = S_IFDIR | 0755; - struct dlmfs_inode_private *ip; + umode_t mode = S_IFDIR | 0755; if (inode) { - ip = DLMFS_I(inode); - inode->i_ino = get_next_ino(); - inode->i_mode = mode; - inode->i_uid = current_fsuid(); - inode->i_gid = current_fsgid(); + inode_init_owner(inode, NULL, mode); inode->i_mapping->backing_dev_info = &dlmfs_backing_dev_info; inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME; inc_nlink(inode); @@ -424,7 +417,7 @@ static struct inode *dlmfs_get_root_inode(struct super_block *sb) static struct inode *dlmfs_get_inode(struct inode *parent, struct dentry *dentry, - int mode) + umode_t mode) { struct super_block *sb = parent->i_sb; struct inode * inode = new_inode(sb); @@ -434,9 +427,7 @@ static struct inode *dlmfs_get_inode(struct inode *parent, return NULL; inode->i_ino = get_next_ino(); - inode->i_mode = mode; - inode->i_uid = current_fsuid(); - inode->i_gid = current_fsgid(); + inode_init_owner(inode, parent, mode); inode->i_mapping->backing_dev_info = &dlmfs_backing_dev_info; inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME; @@ -473,13 +464,6 @@ static struct inode *dlmfs_get_inode(struct inode *parent, inc_nlink(inode); break; } - - if (parent->i_mode & S_ISGID) { - inode->i_gid = parent->i_gid; - if (S_ISDIR(mode)) - inode->i_mode |= S_ISGID; - } - return inode; } @@ -489,7 +473,7 @@ static struct inode *dlmfs_get_inode(struct inode *parent, /* SMP-safe */ static int dlmfs_mkdir(struct inode * dir, struct dentry * dentry, - int mode) + umode_t mode) { int status; struct inode *inode = NULL; @@ -537,8 +521,8 @@ bail: static int dlmfs_create(struct inode *dir, struct dentry *dentry, - int mode, - struct nameidata *nd) + umode_t mode, + bool excl) { int status = 0; struct inode *inode; @@ -594,24 +578,14 @@ static int dlmfs_fill_super(struct super_block * sb, void * data, int silent) { - struct inode * inode; - struct dentry * root; - sb->s_maxbytes = MAX_LFS_FILESIZE; sb->s_blocksize = PAGE_CACHE_SIZE; sb->s_blocksize_bits = PAGE_CACHE_SHIFT; sb->s_magic = DLMFS_MAGIC; sb->s_op = &dlmfs_ops; - inode = dlmfs_get_root_inode(sb); - if (!inode) - return -ENOMEM; - - root = d_alloc_root(inode); - if (!root) { - iput(inode); + sb->s_root = d_make_root(dlmfs_get_root_inode(sb)); + if (!sb->s_root) return -ENOMEM; - } - sb->s_root = root; return 0; } @@ -662,14 +636,13 @@ static struct file_system_type dlmfs_fs_type = { .mount = dlmfs_mount, .kill_sb = kill_litter_super, }; +MODULE_ALIAS_FS("ocfs2_dlmfs"); static int __init init_dlmfs_fs(void) { int status; int cleanup_inode = 0, cleanup_worker = 0; - dlmfs_print_version(); - status = bdi_init(&dlmfs_backing_dev_info); if (status) return status; @@ -713,6 +686,11 @@ static void __exit exit_dlmfs_fs(void) flush_workqueue(user_dlm_worker); destroy_workqueue(user_dlm_worker); + /* + * Make sure all delayed rcu free inodes are flushed before we + * destroy cache. + */ + rcu_barrier(); kmem_cache_destroy(dlmfs_inode_cache); bdi_destroy(&dlmfs_backing_dev_info); @@ -720,6 +698,7 @@ static void __exit exit_dlmfs_fs(void) MODULE_AUTHOR("Oracle"); MODULE_LICENSE("GPL"); +MODULE_DESCRIPTION("OCFS2 DLM-Filesystem"); module_init(init_dlmfs_fs) module_exit(exit_dlmfs_fs) diff --git a/fs/ocfs2/dlmfs/dlmfsver.c b/fs/ocfs2/dlmfs/dlmfsver.c deleted file mode 100644 index a733b3321f8..00000000000 --- a/fs/ocfs2/dlmfs/dlmfsver.c +++ /dev/null @@ -1,42 +0,0 @@ -/* -*- mode: c; c-basic-offset: 8; -*- - * vim: noexpandtab sw=8 ts=8 sts=0: - * - * dlmfsver.c - * - * version string - * - * Copyright (C) 2002, 2005 Oracle. All rights reserved. - * - * This program is free software; you can redistribute it and/or - * modify it under the terms of the GNU General Public - * License as published by the Free Software Foundation; either - * version 2 of the License, or (at your option) any later version. - * - * This program is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * General Public License for more details. - * - * You should have received a copy of the GNU General Public - * License along with this program; if not, write to the - * Free Software Foundation, Inc., 59 Temple Place - Suite 330, - * Boston, MA 021110-1307, USA. - */ - -#include <linux/module.h> -#include <linux/kernel.h> - -#include "dlmfsver.h" - -#define DLM_BUILD_VERSION "1.5.0" - -#define VERSION_STR "OCFS2 DLMFS " DLM_BUILD_VERSION - -void dlmfs_print_version(void) -{ - printk(KERN_INFO "%s\n", VERSION_STR); -} - -MODULE_DESCRIPTION(VERSION_STR); - -MODULE_VERSION(DLM_BUILD_VERSION); diff --git a/fs/ocfs2/dlmfs/dlmfsver.h b/fs/ocfs2/dlmfs/dlmfsver.h deleted file mode 100644 index f35eadbed25..00000000000 --- a/fs/ocfs2/dlmfs/dlmfsver.h +++ /dev/null @@ -1,31 +0,0 @@ -/* -*- mode: c; c-basic-offset: 8; -*- - * vim: noexpandtab sw=8 ts=8 sts=0: - * - * dlmver.h - * - * Function prototypes - * - * Copyright (C) 2005 Oracle. All rights reserved. - * - * This program is free software; you can redistribute it and/or - * modify it under the terms of the GNU General Public - * License as published by the Free Software Foundation; either - * version 2 of the License, or (at your option) any later version. - * - * This program is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * General Public License for more details. - * - * You should have received a copy of the GNU General Public - * License along with this program; if not, write to the - * Free Software Foundation, Inc., 59 Temple Place - Suite 330, - * Boston, MA 021110-1307, USA. - */ - -#ifndef DLMFS_VER_H -#define DLMFS_VER_H - -void dlmfs_print_version(void); - -#endif /* DLMFS_VER_H */ diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c index 7642d7ca73e..52cfe99ae05 100644 --- a/fs/ocfs2/dlmglue.c +++ b/fs/ocfs2/dlmglue.c @@ -456,7 +456,7 @@ static void ocfs2_update_lock_stats(struct ocfs2_lock_res *res, int level, stats->ls_gets++; stats->ls_total += ktime_to_ns(kt); /* overflow */ - if (unlikely(stats->ls_gets) == 0) { + if (unlikely(stats->ls_gets == 0)) { stats->ls_gets++; stats->ls_total = ktime_to_ns(kt); } @@ -1304,7 +1304,7 @@ static int ocfs2_wait_for_mask(struct ocfs2_mask_waiter *mw) { wait_for_completion(&mw->mw_complete); /* Re-arm the completion in case we want to wait on it again */ - INIT_COMPLETION(mw->mw_complete); + reinit_completion(&mw->mw_complete); return mw->mw_status; } @@ -1355,7 +1355,7 @@ static int ocfs2_wait_for_mask_interruptible(struct ocfs2_mask_waiter *mw, else ret = mw->mw_status; /* Re-arm the completion in case we want to wait on it again */ - INIT_COMPLETION(mw->mw_complete); + reinit_completion(&mw->mw_complete); return ret; } @@ -1692,7 +1692,7 @@ int ocfs2_open_lock(struct inode *inode) mlog(0, "inode %llu take PRMODE open lock\n", (unsigned long long)OCFS2_I(inode)->ip_blkno); - if (ocfs2_mount_local(osb)) + if (ocfs2_is_hard_readonly(osb) || ocfs2_mount_local(osb)) goto out; lockres = &OCFS2_I(inode)->ip_open_lockres; @@ -1718,6 +1718,12 @@ int ocfs2_try_open_lock(struct inode *inode, int write) (unsigned long long)OCFS2_I(inode)->ip_blkno, write ? "EXMODE" : "PRMODE"); + if (ocfs2_is_hard_readonly(osb)) { + if (write) + status = -EROFS; + goto out; + } + if (ocfs2_mount_local(osb)) goto out; @@ -2039,8 +2045,8 @@ static void __ocfs2_stuff_meta_lvb(struct inode *inode) lvb->lvb_version = OCFS2_LVB_VERSION; lvb->lvb_isize = cpu_to_be64(i_size_read(inode)); lvb->lvb_iclusters = cpu_to_be32(oi->ip_clusters); - lvb->lvb_iuid = cpu_to_be32(inode->i_uid); - lvb->lvb_igid = cpu_to_be32(inode->i_gid); + lvb->lvb_iuid = cpu_to_be32(i_uid_read(inode)); + lvb->lvb_igid = cpu_to_be32(i_gid_read(inode)); lvb->lvb_imode = cpu_to_be16(inode->i_mode); lvb->lvb_inlink = cpu_to_be16(inode->i_nlink); lvb->lvb_iatime_packed = @@ -2089,10 +2095,10 @@ static void ocfs2_refresh_inode_from_lvb(struct inode *inode) else inode->i_blocks = ocfs2_inode_sector_count(inode); - inode->i_uid = be32_to_cpu(lvb->lvb_iuid); - inode->i_gid = be32_to_cpu(lvb->lvb_igid); + i_uid_write(inode, be32_to_cpu(lvb->lvb_iuid)); + i_gid_write(inode, be32_to_cpu(lvb->lvb_igid)); inode->i_mode = be16_to_cpu(lvb->lvb_imode); - inode->i_nlink = be16_to_cpu(lvb->lvb_inlink); + set_nlink(inode, be16_to_cpu(lvb->lvb_inlink)); ocfs2_unpack_timespec(&inode->i_atime, be64_to_cpu(lvb->lvb_iatime_packed)); ocfs2_unpack_timespec(&inode->i_mtime, @@ -2298,7 +2304,7 @@ int ocfs2_inode_lock_full_nested(struct inode *inode, if (ocfs2_is_hard_readonly(osb)) { if (ex) status = -EROFS; - goto bail; + goto getbh; } if (ocfs2_mount_local(osb)) @@ -2316,7 +2322,7 @@ int ocfs2_inode_lock_full_nested(struct inode *inode, status = __ocfs2_cluster_lock(osb, lockres, level, dlm_flags, arg_flags, subclass, _RET_IP_); if (status < 0) { - if (status != -EAGAIN && status != -EIOCBRETRY) + if (status != -EAGAIN) mlog_errno(status); goto bail; } @@ -2356,7 +2362,7 @@ local: mlog_errno(status); goto bail; } - +getbh: if (ret_bh) { status = ocfs2_assign_bh(inode, ret_bh, local_bh); if (status < 0) { @@ -2538,17 +2544,15 @@ int ocfs2_super_lock(struct ocfs2_super *osb, * refreshed, so we do it here. Of course, making sense of * everything is up to the caller :) */ status = ocfs2_should_refresh_lock_res(lockres); - if (status < 0) { - mlog_errno(status); - goto bail; - } if (status) { status = ocfs2_refresh_slot_info(osb); ocfs2_complete_lock_res_refresh(lockres, status); - if (status < 0) + if (status < 0) { + ocfs2_cluster_unlock(osb, lockres, level); mlog_errno(status); + } ocfs2_track_lock_refresh(lockres); } bail: @@ -2628,8 +2632,11 @@ int ocfs2_dentry_lock(struct dentry *dentry, int ex) BUG_ON(!dl); - if (ocfs2_is_hard_readonly(osb)) - return -EROFS; + if (ocfs2_is_hard_readonly(osb)) { + if (ex) + return -EROFS; + return 0; + } if (ocfs2_mount_local(osb)) return 0; @@ -2647,7 +2654,7 @@ void ocfs2_dentry_unlock(struct dentry *dentry, int ex) struct ocfs2_dentry_lock *dl = dentry->d_fsdata; struct ocfs2_super *osb = OCFS2_SB(dentry->d_sb); - if (!ocfs2_mount_local(osb)) + if (!ocfs2_is_hard_readonly(osb) && !ocfs2_mount_local(osb)) ocfs2_cluster_unlock(osb, &dl->dl_lockres, level); } @@ -2984,6 +2991,8 @@ int ocfs2_dlm_init(struct ocfs2_super *osb) /* for now, uuid == domain */ status = ocfs2_cluster_connect(osb->osb_cluster_stack, + osb->osb_cluster_name, + strlen(osb->osb_cluster_name), osb->uuid_str, strlen(osb->uuid_str), &lproto, ocfs2_do_node_down, osb, @@ -2993,7 +3002,7 @@ int ocfs2_dlm_init(struct ocfs2_super *osb) goto bail; } - status = ocfs2_cluster_this_node(&osb->node_num); + status = ocfs2_cluster_this_node(conn, &osb->node_num); if (status < 0) { mlog_errno(status); mlog(ML_ERROR, @@ -3130,22 +3139,60 @@ out: return 0; } +static void ocfs2_process_blocked_lock(struct ocfs2_super *osb, + struct ocfs2_lock_res *lockres); + /* Mark the lockres as being dropped. It will no longer be * queued if blocking, but we still may have to wait on it * being dequeued from the downconvert thread before we can consider * it safe to drop. * * You can *not* attempt to call cluster_lock on this lockres anymore. */ -void ocfs2_mark_lockres_freeing(struct ocfs2_lock_res *lockres) +void ocfs2_mark_lockres_freeing(struct ocfs2_super *osb, + struct ocfs2_lock_res *lockres) { int status; struct ocfs2_mask_waiter mw; - unsigned long flags; + unsigned long flags, flags2; ocfs2_init_mask_waiter(&mw); spin_lock_irqsave(&lockres->l_lock, flags); lockres->l_flags |= OCFS2_LOCK_FREEING; + if (lockres->l_flags & OCFS2_LOCK_QUEUED && current == osb->dc_task) { + /* + * We know the downconvert is queued but not in progress + * because we are the downconvert thread and processing + * different lock. So we can just remove the lock from the + * queue. This is not only an optimization but also a way + * to avoid the following deadlock: + * ocfs2_dentry_post_unlock() + * ocfs2_dentry_lock_put() + * ocfs2_drop_dentry_lock() + * iput() + * ocfs2_evict_inode() + * ocfs2_clear_inode() + * ocfs2_mark_lockres_freeing() + * ... blocks waiting for OCFS2_LOCK_QUEUED + * since we are the downconvert thread which + * should clear the flag. + */ + spin_unlock_irqrestore(&lockres->l_lock, flags); + spin_lock_irqsave(&osb->dc_task_lock, flags2); + list_del_init(&lockres->l_blocked_list); + osb->blocked_lock_count--; + spin_unlock_irqrestore(&osb->dc_task_lock, flags2); + /* + * Warn if we recurse into another post_unlock call. Strictly + * speaking it isn't a problem but we need to be careful if + * that happens (stack overflow, deadlocks, ...) so warn if + * ocfs2 grows a path for which this can happen. + */ + WARN_ON_ONCE(lockres->l_ops->post_unlock); + /* Since the lock is freeing we don't do much in the fn below */ + ocfs2_process_blocked_lock(osb, lockres); + return; + } while (lockres->l_flags & OCFS2_LOCK_QUEUED) { lockres_add_mask_waiter(lockres, &mw, OCFS2_LOCK_QUEUED, 0); spin_unlock_irqrestore(&lockres->l_lock, flags); @@ -3166,7 +3213,7 @@ void ocfs2_simple_drop_lockres(struct ocfs2_super *osb, { int ret; - ocfs2_mark_lockres_freeing(lockres); + ocfs2_mark_lockres_freeing(osb, lockres); ret = ocfs2_drop_lock(osb, lockres); if (ret) mlog_errno(ret); @@ -3923,6 +3970,8 @@ unqueue: static void ocfs2_schedule_blocked_lock(struct ocfs2_super *osb, struct ocfs2_lock_res *lockres) { + unsigned long flags; + assert_spin_locked(&lockres->l_lock); if (lockres->l_flags & OCFS2_LOCK_FREEING) { @@ -3936,21 +3985,22 @@ static void ocfs2_schedule_blocked_lock(struct ocfs2_super *osb, lockres_or_flags(lockres, OCFS2_LOCK_QUEUED); - spin_lock(&osb->dc_task_lock); + spin_lock_irqsave(&osb->dc_task_lock, flags); if (list_empty(&lockres->l_blocked_list)) { list_add_tail(&lockres->l_blocked_list, &osb->blocked_lock_list); osb->blocked_lock_count++; } - spin_unlock(&osb->dc_task_lock); + spin_unlock_irqrestore(&osb->dc_task_lock, flags); } static void ocfs2_downconvert_thread_do_work(struct ocfs2_super *osb) { unsigned long processed; + unsigned long flags; struct ocfs2_lock_res *lockres; - spin_lock(&osb->dc_task_lock); + spin_lock_irqsave(&osb->dc_task_lock, flags); /* grab this early so we know to try again if a state change and * wake happens part-way through our work */ osb->dc_work_sequence = osb->dc_wake_sequence; @@ -3963,38 +4013,40 @@ static void ocfs2_downconvert_thread_do_work(struct ocfs2_super *osb) struct ocfs2_lock_res, l_blocked_list); list_del_init(&lockres->l_blocked_list); osb->blocked_lock_count--; - spin_unlock(&osb->dc_task_lock); + spin_unlock_irqrestore(&osb->dc_task_lock, flags); BUG_ON(!processed); processed--; ocfs2_process_blocked_lock(osb, lockres); - spin_lock(&osb->dc_task_lock); + spin_lock_irqsave(&osb->dc_task_lock, flags); } - spin_unlock(&osb->dc_task_lock); + spin_unlock_irqrestore(&osb->dc_task_lock, flags); } static int ocfs2_downconvert_thread_lists_empty(struct ocfs2_super *osb) { int empty = 0; + unsigned long flags; - spin_lock(&osb->dc_task_lock); + spin_lock_irqsave(&osb->dc_task_lock, flags); if (list_empty(&osb->blocked_lock_list)) empty = 1; - spin_unlock(&osb->dc_task_lock); + spin_unlock_irqrestore(&osb->dc_task_lock, flags); return empty; } static int ocfs2_downconvert_thread_should_wake(struct ocfs2_super *osb) { int should_wake = 0; + unsigned long flags; - spin_lock(&osb->dc_task_lock); + spin_lock_irqsave(&osb->dc_task_lock, flags); if (osb->dc_work_sequence != osb->dc_wake_sequence) should_wake = 1; - spin_unlock(&osb->dc_task_lock); + spin_unlock_irqrestore(&osb->dc_task_lock, flags); return should_wake; } @@ -4024,10 +4076,12 @@ static int ocfs2_downconvert_thread(void *arg) void ocfs2_wake_downconvert_thread(struct ocfs2_super *osb) { - spin_lock(&osb->dc_task_lock); + unsigned long flags; + + spin_lock_irqsave(&osb->dc_task_lock, flags); /* make sure the voting thread gets a swipe at whatever changes * the caller may have made to the voting state */ osb->dc_wake_sequence++; - spin_unlock(&osb->dc_task_lock); + spin_unlock_irqrestore(&osb->dc_task_lock, flags); wake_up(&osb->dc_event); } diff --git a/fs/ocfs2/dlmglue.h b/fs/ocfs2/dlmglue.h index 1d596d8c4a4..d293a22c32c 100644 --- a/fs/ocfs2/dlmglue.h +++ b/fs/ocfs2/dlmglue.h @@ -157,7 +157,8 @@ int ocfs2_refcount_lock(struct ocfs2_refcount_tree *ref_tree, int ex); void ocfs2_refcount_unlock(struct ocfs2_refcount_tree *ref_tree, int ex); -void ocfs2_mark_lockres_freeing(struct ocfs2_lock_res *lockres); +void ocfs2_mark_lockres_freeing(struct ocfs2_super *osb, + struct ocfs2_lock_res *lockres); void ocfs2_simple_drop_lockres(struct ocfs2_super *osb, struct ocfs2_lock_res *lockres); diff --git a/fs/ocfs2/export.c b/fs/ocfs2/export.c index 745db42528d..29651167190 100644 --- a/fs/ocfs2/export.c +++ b/fs/ocfs2/export.c @@ -177,27 +177,29 @@ bail: return parent; } -static int ocfs2_encode_fh(struct dentry *dentry, u32 *fh_in, int *max_len, - int connectable) +static int ocfs2_encode_fh(struct inode *inode, u32 *fh_in, int *max_len, + struct inode *parent) { - struct inode *inode = dentry->d_inode; int len = *max_len; int type = 1; u64 blkno; u32 generation; __le32 *fh = (__force __le32 *) fh_in; +#ifdef TRACE_HOOKS_ARE_NOT_BRAINDEAD_IN_YOUR_OPINION +#error "You go ahead and fix that mess, then. Somehow" trace_ocfs2_encode_fh_begin(dentry, dentry->d_name.len, dentry->d_name.name, fh, len, connectable); +#endif - if (connectable && (len < 6)) { + if (parent && (len < 6)) { *max_len = 6; - type = 255; + type = FILEID_INVALID; goto bail; } else if (len < 3) { *max_len = 3; - type = 255; + type = FILEID_INVALID; goto bail; } @@ -211,12 +213,7 @@ static int ocfs2_encode_fh(struct dentry *dentry, u32 *fh_in, int *max_len, fh[1] = cpu_to_le32((u32)(blkno & 0xffffffff)); fh[2] = cpu_to_le32(generation); - if (connectable && !S_ISDIR(inode->i_mode)) { - struct inode *parent; - - spin_lock(&dentry->d_lock); - - parent = dentry->d_parent->d_inode; + if (parent) { blkno = OCFS2_I(parent)->ip_blkno; generation = parent->i_generation; @@ -224,8 +221,6 @@ static int ocfs2_encode_fh(struct dentry *dentry, u32 *fh_in, int *max_len, fh[4] = cpu_to_le32((u32)(blkno & 0xffffffff)); fh[5] = cpu_to_le32(generation); - spin_unlock(&dentry->d_lock); - len = 6; type = 2; diff --git a/fs/ocfs2/extent_map.c b/fs/ocfs2/extent_map.c index 23457b491e8..767370b656c 100644 --- a/fs/ocfs2/extent_map.c +++ b/fs/ocfs2/extent_map.c @@ -282,8 +282,7 @@ search: spin_unlock(&oi->ip_lock); out: - if (new_emi) - kfree(new_emi); + kfree(new_emi); } static int ocfs2_last_eb_is_empty(struct inode *inode, @@ -782,7 +781,6 @@ int ocfs2_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, cpos = map_start >> osb->s_clustersize_bits; mapping_end = ocfs2_clusters_for_bytes(inode->i_sb, map_start + map_len); - mapping_end -= cpos; is_last = 0; while (cpos < mapping_end && !is_last) { u32 fe_flags; @@ -791,7 +789,7 @@ int ocfs2_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, &hole_size, &rec, &is_last); if (ret) { mlog_errno(ret); - goto out; + goto out_unlock; } if (rec.e_blkno == 0ULL) { @@ -832,6 +830,100 @@ out: return ret; } +int ocfs2_seek_data_hole_offset(struct file *file, loff_t *offset, int whence) +{ + struct inode *inode = file->f_mapping->host; + int ret; + unsigned int is_last = 0, is_data = 0; + u16 cs_bits = OCFS2_SB(inode->i_sb)->s_clustersize_bits; + u32 cpos, cend, clen, hole_size; + u64 extoff, extlen; + struct buffer_head *di_bh = NULL; + struct ocfs2_extent_rec rec; + + BUG_ON(whence != SEEK_DATA && whence != SEEK_HOLE); + + ret = ocfs2_inode_lock(inode, &di_bh, 0); + if (ret) { + mlog_errno(ret); + goto out; + } + + down_read(&OCFS2_I(inode)->ip_alloc_sem); + + if (*offset >= i_size_read(inode)) { + ret = -ENXIO; + goto out_unlock; + } + + if (OCFS2_I(inode)->ip_dyn_features & OCFS2_INLINE_DATA_FL) { + if (whence == SEEK_HOLE) + *offset = i_size_read(inode); + goto out_unlock; + } + + clen = 0; + cpos = *offset >> cs_bits; + cend = ocfs2_clusters_for_bytes(inode->i_sb, i_size_read(inode)); + + while (cpos < cend && !is_last) { + ret = ocfs2_get_clusters_nocache(inode, di_bh, cpos, &hole_size, + &rec, &is_last); + if (ret) { + mlog_errno(ret); + goto out_unlock; + } + + extoff = cpos; + extoff <<= cs_bits; + + if (rec.e_blkno == 0ULL) { + clen = hole_size; + is_data = 0; + } else { + clen = le16_to_cpu(rec.e_leaf_clusters) - + (cpos - le32_to_cpu(rec.e_cpos)); + is_data = (rec.e_flags & OCFS2_EXT_UNWRITTEN) ? 0 : 1; + } + + if ((!is_data && whence == SEEK_HOLE) || + (is_data && whence == SEEK_DATA)) { + if (extoff > *offset) + *offset = extoff; + goto out_unlock; + } + + if (!is_last) + cpos += clen; + } + + if (whence == SEEK_HOLE) { + extoff = cpos; + extoff <<= cs_bits; + extlen = clen; + extlen <<= cs_bits; + + if ((extoff + extlen) > i_size_read(inode)) + extlen = i_size_read(inode) - extoff; + extoff += extlen; + if (extoff > *offset) + *offset = extoff; + goto out_unlock; + } + + ret = -ENXIO; + +out_unlock: + + brelse(di_bh); + + up_read(&OCFS2_I(inode)->ip_alloc_sem); + + ocfs2_inode_unlock(inode, 0); +out: + return ret; +} + int ocfs2_read_virt_blocks(struct inode *inode, u64 v_block, int nr, struct buffer_head *bhs[], int flags, int (*validate)(struct super_block *sb, diff --git a/fs/ocfs2/extent_map.h b/fs/ocfs2/extent_map.h index e79d41c2c90..67ea57d2fd5 100644 --- a/fs/ocfs2/extent_map.h +++ b/fs/ocfs2/extent_map.h @@ -53,6 +53,8 @@ int ocfs2_extent_map_get_blocks(struct inode *inode, u64 v_blkno, u64 *p_blkno, int ocfs2_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, u64 map_start, u64 map_len); +int ocfs2_seek_data_hole_offset(struct file *file, loff_t *offset, int origin); + int ocfs2_xattr_get_clusters(struct inode *inode, u32 v_cluster, u32 *p_cluster, u32 *num_clusters, struct ocfs2_extent_list *el, diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c index 41565ae5285..2930e231f3f 100644 --- a/fs/ocfs2/file.c +++ b/fs/ocfs2/file.c @@ -171,12 +171,17 @@ static int ocfs2_dir_release(struct inode *inode, struct file *file) return 0; } -static int ocfs2_sync_file(struct file *file, int datasync) +static int ocfs2_sync_file(struct file *file, loff_t start, loff_t end, + int datasync) { int err = 0; - journal_t *journal; struct inode *inode = file->f_mapping->host; struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); + struct ocfs2_inode_info *oi = OCFS2_I(inode); + journal_t *journal = osb->journal->j_journal; + int ret; + tid_t commit_tid; + bool needs_barrier = false; trace_ocfs2_sync_file(inode, file, file->f_path.dentry, OCFS2_I(inode)->ip_blkno, @@ -184,20 +189,24 @@ static int ocfs2_sync_file(struct file *file, int datasync) file->f_path.dentry->d_name.name, (unsigned long long)datasync); - if (datasync && !(inode->i_state & I_DIRTY_DATASYNC)) { - /* - * We still have to flush drive's caches to get data to the - * platter - */ - if (osb->s_mount_opt & OCFS2_MOUNT_BARRIER) - blkdev_issue_flush(inode->i_sb->s_bdev, GFP_KERNEL, NULL); - goto bail; - } + if (ocfs2_is_hard_readonly(osb) || ocfs2_is_soft_readonly(osb)) + return -EROFS; + + err = filemap_write_and_wait_range(inode->i_mapping, start, end); + if (err) + return err; - journal = osb->journal->j_journal; - err = jbd2_journal_force_commit(journal); + commit_tid = datasync ? oi->i_datasync_tid : oi->i_sync_tid; + if (journal->j_flags & JBD2_BARRIER && + !jbd2_trans_will_send_data_barrier(journal, commit_tid)) + needs_barrier = true; + err = jbd2_complete_transaction(journal, commit_tid); + if (needs_barrier) { + ret = blkdev_issue_flush(inode->i_sb->s_bdev, GFP_KERNEL, NULL); + if (!err) + err = ret; + } -bail: if (err) mlog_errno(err); @@ -277,6 +286,7 @@ int ocfs2_update_inode_atime(struct inode *inode, inode->i_atime = CURRENT_TIME; di->i_atime = cpu_to_le64(inode->i_atime.tv_sec); di->i_atime_nsec = cpu_to_le32(inode->i_atime.tv_nsec); + ocfs2_update_inode_fsync_trans(handle, inode, 0); ocfs2_journal_dirty(handle, bh); out_commit: @@ -326,6 +336,7 @@ int ocfs2_simple_size_update(struct inode *inode, if (ret < 0) mlog_errno(ret); + ocfs2_update_inode_fsync_trans(handle, inode, 0); ocfs2_commit_trans(osb, handle); out: return ret; @@ -358,7 +369,7 @@ static int ocfs2_cow_file_pos(struct inode *inode, if (!(ext_flags & OCFS2_EXT_REFCOUNTED)) goto out; - return ocfs2_refcount_cow(inode, NULL, fe_bh, cpos, 1, cpos+1); + return ocfs2_refcount_cow(inode, fe_bh, cpos, 1, cpos+1); out: return status; @@ -420,6 +431,7 @@ static int ocfs2_orphan_for_truncate(struct ocfs2_super *osb, di->i_size = cpu_to_le64(new_i_size); di->i_ctime = di->i_mtime = cpu_to_le64(inode->i_ctime.tv_sec); di->i_ctime_nsec = di->i_mtime_nsec = cpu_to_le32(inode->i_ctime.tv_nsec); + ocfs2_update_inode_fsync_trans(handle, inode, 0); ocfs2_journal_dirty(handle, fe_bh); @@ -462,11 +474,6 @@ static int ocfs2_truncate_file(struct inode *inode, goto bail; } - /* lets handle the simple truncate cases before doing any more - * cluster locking. */ - if (new_i_size == le64_to_cpu(fe->i_size)) - goto bail; - down_write(&OCFS2_I(inode)->ip_alloc_sem); ocfs2_resv_discard(&osb->osb_la_resmap, @@ -568,7 +575,7 @@ static int __ocfs2_extend_allocation(struct inode *inode, u32 logical_start, int did_quota = 0; /* - * This function only exists for file systems which don't + * Unwritten extent only exists for file systems which * support holes. */ BUG_ON(mark_unwritten && !ocfs2_sparse_alloc(osb)); @@ -591,8 +598,7 @@ restart_all: goto leave; } - credits = ocfs2_calc_extend_credits(osb->sb, &fe->id2.i_list, - clusters_to_add); + credits = ocfs2_calc_extend_credits(osb->sb, &fe->id2.i_list); handle = ocfs2_start_trans(osb, credits); if (IS_ERR(handle)) { status = PTR_ERR(handle); @@ -641,7 +647,7 @@ restarted_transaction: mlog_errno(status); goto leave; } - + ocfs2_update_inode_fsync_trans(handle, inode, 1); ocfs2_journal_dirty(handle, bh); spin_lock(&OCFS2_I(inode)->ip_lock); @@ -659,11 +665,7 @@ restarted_transaction: } else { BUG_ON(why != RESTART_TRANS); - /* TODO: This can be more intelligent. */ - credits = ocfs2_calc_extend_credits(osb->sb, - &fe->id2.i_list, - clusters_to_add); - status = ocfs2_extend_trans(handle, credits); + status = ocfs2_allocate_extend_trans(handle, 1); if (status < 0) { /* handle still has to be committed at * this point. */ @@ -711,7 +713,8 @@ leave: * While a write will already be ordering the data, a truncate will not. * Thus, we need to explicitly order the zeroed pages. */ -static handle_t *ocfs2_zero_start_ordered_transaction(struct inode *inode) +static handle_t *ocfs2_zero_start_ordered_transaction(struct inode *inode, + struct buffer_head *di_bh) { struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); handle_t *handle = NULL; @@ -728,8 +731,16 @@ static handle_t *ocfs2_zero_start_ordered_transaction(struct inode *inode) } ret = ocfs2_jbd2_file_inode(handle, inode); - if (ret < 0) + if (ret < 0) { mlog_errno(ret); + goto out; + } + + ret = ocfs2_journal_access_di(handle, INODE_CACHE(inode), di_bh, + OCFS2_JOURNAL_ACCESS_WRITE); + if (ret) + mlog_errno(ret); + ocfs2_update_inode_fsync_trans(handle, inode, 1); out: if (ret) { @@ -744,7 +755,7 @@ out: * to be too fragile to do exactly what we need without us having to * worry about recursive locking in ->write_begin() and ->write_end(). */ static int ocfs2_write_zero_page(struct inode *inode, u64 abs_from, - u64 abs_to) + u64 abs_to, struct buffer_head *di_bh) { struct address_space *mapping = inode->i_mapping; struct page *page; @@ -752,6 +763,7 @@ static int ocfs2_write_zero_page(struct inode *inode, u64 abs_from, handle_t *handle = NULL; int ret = 0; unsigned zero_from, zero_to, block_start, block_end; + struct ocfs2_dinode *di = (struct ocfs2_dinode *)di_bh->b_data; BUG_ON(abs_from >= abs_to); BUG_ON(abs_to > (((u64)index + 1) << PAGE_CACHE_SHIFT)); @@ -794,7 +806,8 @@ static int ocfs2_write_zero_page(struct inode *inode, u64 abs_from, } if (!handle) { - handle = ocfs2_zero_start_ordered_transaction(inode); + handle = ocfs2_zero_start_ordered_transaction(inode, + di_bh); if (IS_ERR(handle)) { ret = PTR_ERR(handle); handle = NULL; @@ -811,8 +824,23 @@ static int ocfs2_write_zero_page(struct inode *inode, u64 abs_from, ret = 0; } - if (handle) + if (handle) { + /* + * fs-writeback will release the dirty pages without page lock + * whose offset are over inode size, the release happens at + * block_write_full_page(). + */ + i_size_write(inode, abs_to); + inode->i_blocks = ocfs2_inode_sector_count(inode); + di->i_size = cpu_to_le64((u64)i_size_read(inode)); + inode->i_mtime = inode->i_ctime = CURRENT_TIME; + di->i_mtime = di->i_ctime = cpu_to_le64(inode->i_mtime.tv_sec); + di->i_ctime_nsec = cpu_to_le32(inode->i_mtime.tv_nsec); + di->i_mtime_nsec = di->i_ctime_nsec; + ocfs2_journal_dirty(handle, di_bh); + ocfs2_update_inode_fsync_trans(handle, inode, 1); ocfs2_commit_trans(OCFS2_SB(inode->i_sb), handle); + } out_unlock: unlock_page(page); @@ -887,7 +915,7 @@ static int ocfs2_zero_extend_get_range(struct inode *inode, zero_clusters = last_cpos - zero_cpos; if (needs_cow) { - rc = ocfs2_refcount_cow(inode, NULL, di_bh, zero_cpos, + rc = ocfs2_refcount_cow(inode, di_bh, zero_cpos, zero_clusters, UINT_MAX); if (rc) { mlog_errno(rc); @@ -908,7 +936,7 @@ out: * has made sure that the entire range needs zeroing. */ static int ocfs2_zero_extend_range(struct inode *inode, u64 range_start, - u64 range_end) + u64 range_end, struct buffer_head *di_bh) { int rc = 0; u64 next_pos; @@ -924,7 +952,7 @@ static int ocfs2_zero_extend_range(struct inode *inode, u64 range_start, next_pos = (zero_pos & PAGE_CACHE_MASK) + PAGE_CACHE_SIZE; if (next_pos > range_end) next_pos = range_end; - rc = ocfs2_write_zero_page(inode, zero_pos, next_pos); + rc = ocfs2_write_zero_page(inode, zero_pos, next_pos, di_bh); if (rc < 0) { mlog_errno(rc); break; @@ -970,7 +998,7 @@ int ocfs2_zero_extend(struct inode *inode, struct buffer_head *di_bh, range_end = zero_to_size; ret = ocfs2_zero_extend_range(inode, range_start, - range_end); + range_end, di_bh); if (ret) { mlog_errno(ret); break; @@ -1104,7 +1132,8 @@ int ocfs2_setattr(struct dentry *dentry, struct iattr *attr) (unsigned long long)OCFS2_I(inode)->ip_blkno, dentry->d_name.len, dentry->d_name.name, attr->ia_valid, attr->ia_mode, - attr->ia_uid, attr->ia_gid); + from_kuid(&init_user_ns, attr->ia_uid), + from_kgid(&init_user_ns, attr->ia_gid)); /* ensuring we don't even attempt to truncate a symlink */ if (S_ISLNK(inode->i_mode)) @@ -1137,12 +1166,14 @@ int ocfs2_setattr(struct dentry *dentry, struct iattr *attr) goto bail_unlock_rw; } - if (size_change && attr->ia_size != i_size_read(inode)) { + if (size_change) { status = inode_newsize_ok(inode, attr->ia_size); if (status) goto bail_unlock; - if (i_size_read(inode) > attr->ia_size) { + inode_dio_wait(inode); + + if (i_size_read(inode) >= attr->ia_size) { if (ocfs2_should_order_data(inode)) { status = ocfs2_begin_ordered_truncate(inode, attr->ia_size); @@ -1160,28 +1191,26 @@ int ocfs2_setattr(struct dentry *dentry, struct iattr *attr) } } - if ((attr->ia_valid & ATTR_UID && attr->ia_uid != inode->i_uid) || - (attr->ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid)) { + if ((attr->ia_valid & ATTR_UID && !uid_eq(attr->ia_uid, inode->i_uid)) || + (attr->ia_valid & ATTR_GID && !gid_eq(attr->ia_gid, inode->i_gid))) { /* * Gather pointers to quota structures so that allocation / * freeing of quota structures happens here and not inside * dquot_transfer() where we have problems with lock ordering */ - if (attr->ia_valid & ATTR_UID && attr->ia_uid != inode->i_uid + if (attr->ia_valid & ATTR_UID && !uid_eq(attr->ia_uid, inode->i_uid) && OCFS2_HAS_RO_COMPAT_FEATURE(sb, OCFS2_FEATURE_RO_COMPAT_USRQUOTA)) { - transfer_to[USRQUOTA] = dqget(sb, attr->ia_uid, - USRQUOTA); + transfer_to[USRQUOTA] = dqget(sb, make_kqid_uid(attr->ia_uid)); if (!transfer_to[USRQUOTA]) { status = -ESRCH; goto bail_unlock; } } - if (attr->ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid + if (attr->ia_valid & ATTR_GID && !gid_eq(attr->ia_gid, inode->i_gid) && OCFS2_HAS_RO_COMPAT_FEATURE(sb, OCFS2_FEATURE_RO_COMPAT_GRPQUOTA)) { - transfer_to[GRPQUOTA] = dqget(sb, attr->ia_gid, - GRPQUOTA); + transfer_to[GRPQUOTA] = dqget(sb, make_kqid_gid(attr->ia_gid)); if (!transfer_to[GRPQUOTA]) { status = -ESRCH; goto bail_unlock; @@ -1206,24 +1235,6 @@ int ocfs2_setattr(struct dentry *dentry, struct iattr *attr) } } - /* - * This will intentionally not wind up calling truncate_setsize(), - * since all the work for a size change has been done above. - * Otherwise, we could get into problems with truncate as - * ip_alloc_sem is used there to protect against i_size - * changes. - * - * XXX: this means the conditional below can probably be removed. - */ - if ((attr->ia_valid & ATTR_SIZE) && - attr->ia_size != i_size_read(inode)) { - status = vmtruncate(inode, attr->ia_size); - if (status) { - mlog_errno(status); - goto bail_commit; - } - } - setattr_copy(inode, attr); mark_inode_dirty(inode); @@ -1246,7 +1257,7 @@ bail: dqput(transfer_to[qtype]); if (!status && attr->ia_valid & ATTR_MODE) { - status = ocfs2_acl_chmod(inode); + status = posix_acl_chmod(inode, inode->i_mode); if (status < 0) mlog_errno(status); } @@ -1279,11 +1290,11 @@ bail: return err; } -int ocfs2_permission(struct inode *inode, int mask, unsigned int flags) +int ocfs2_permission(struct inode *inode, int mask) { int ret; - if (flags & IPERM_FLAG_RCU) + if (mask & MAY_NOT_BLOCK) return -ECHILD; ret = ocfs2_inode_lock(inode, NULL, 0); @@ -1293,7 +1304,7 @@ int ocfs2_permission(struct inode *inode, int mask, unsigned int flags) goto out; } - ret = generic_permission(inode, mask, flags, ocfs2_check_acl); + ret = generic_permission(inode, mask); ocfs2_inode_unlock(inode, 0); out: @@ -1332,6 +1343,7 @@ static int __ocfs2_write_remove_suid(struct inode *inode, di = (struct ocfs2_dinode *) bh->b_data; di->i_mode = cpu_to_le16(inode->i_mode); + ocfs2_update_inode_fsync_trans(handle, inode, 0); ocfs2_journal_dirty(handle, bh); @@ -1564,6 +1576,7 @@ static int ocfs2_zero_partial_clusters(struct inode *inode, if (ret) mlog_errno(ret); } + ocfs2_update_inode_fsync_trans(handle, inode, 1); ocfs2_commit_trans(osb, handle); out: @@ -1607,6 +1620,9 @@ static void ocfs2_calc_trunc_pos(struct inode *inode, range = le32_to_cpu(rec->e_cpos) + ocfs2_rec_clusters(el, rec); if (le32_to_cpu(rec->e_cpos) >= trunc_start) { + /* + * remove an entire extent record. + */ *trunc_cpos = le32_to_cpu(rec->e_cpos); /* * Skip holes if any. @@ -1617,7 +1633,16 @@ static void ocfs2_calc_trunc_pos(struct inode *inode, *blkno = le64_to_cpu(rec->e_blkno); *trunc_end = le32_to_cpu(rec->e_cpos); } else if (range > trunc_start) { + /* + * remove a partial extent record, which means we're + * removing the last extent record. + */ *trunc_cpos = trunc_start; + /* + * skip hole if any. + */ + if (range < *trunc_end) + *trunc_end = range; *trunc_len = *trunc_end - trunc_start; coff = trunc_start - le32_to_cpu(rec->e_cpos); *blkno = le64_to_cpu(rec->e_blkno) + @@ -1793,6 +1818,7 @@ static int ocfs2_remove_inode_range(struct inode *inode, ocfs2_truncate_cluster_pages(inode, byte_start, byte_len); out: + ocfs2_free_path(path); ocfs2_schedule_truncate_log_flush(osb, 1); ocfs2_run_deallocs(osb, &dealloc); @@ -1866,7 +1892,8 @@ static int __ocfs2_change_file_space(struct file *file, struct inode *inode, } size = sr->l_start + sr->l_len; - if (cmd == OCFS2_IOC_RESVSP || cmd == OCFS2_IOC_RESVSP64) { + if (cmd == OCFS2_IOC_RESVSP || cmd == OCFS2_IOC_RESVSP64 || + cmd == OCFS2_IOC_UNRESVSP || cmd == OCFS2_IOC_UNRESVSP64) { if (sr->l_len <= 0) { ret = -EINVAL; goto out_inode_unlock; @@ -1924,6 +1951,9 @@ static int __ocfs2_change_file_space(struct file *file, struct inode *inode, if (ret < 0) mlog_errno(ret); + if (file && (file->f_flags & O_SYNC)) + handle->h_sync = 1; + ocfs2_commit_trans(osb, handle); out_inode_unlock: @@ -1940,8 +1970,9 @@ out: int ocfs2_change_file_space(struct file *file, unsigned int cmd, struct ocfs2_space_resv *sr) { - struct inode *inode = file->f_path.dentry->d_inode; + struct inode *inode = file_inode(file); struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); + int ret; if ((cmd == OCFS2_IOC_RESVSP || cmd == OCFS2_IOC_RESVSP64) && !ocfs2_writes_unwritten_extents(osb)) @@ -1956,13 +1987,18 @@ int ocfs2_change_file_space(struct file *file, unsigned int cmd, if (!(file->f_mode & FMODE_WRITE)) return -EBADF; - return __ocfs2_change_file_space(file, inode, file->f_pos, cmd, sr, 0); + ret = mnt_want_write_file(file); + if (ret) + return ret; + ret = __ocfs2_change_file_space(file, inode, file->f_pos, cmd, sr, 0); + mnt_drop_write_file(file); + return ret; } static long ocfs2_fallocate(struct file *file, int mode, loff_t offset, loff_t len) { - struct inode *inode = file->f_path.dentry->d_inode; + struct inode *inode = file_inode(file); struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); struct ocfs2_space_resv sr; int change_size = 1; @@ -2026,6 +2062,16 @@ out: return ret; } +static int ocfs2_is_io_unaligned(struct inode *inode, size_t count, loff_t pos) +{ + int blockmask = inode->i_sb->s_blocksize - 1; + loff_t final_size = pos + count; + + if ((pos & blockmask) || (final_size & blockmask)) + return 1; + return 0; +} + static int ocfs2_prepare_inode_for_refcount(struct inode *inode, struct file *file, loff_t pos, size_t count, @@ -2045,7 +2091,7 @@ static int ocfs2_prepare_inode_for_refcount(struct inode *inode, *meta_level = 1; - ret = ocfs2_refcount_cow(inode, file, di_bh, cpos, clusters, UINT_MAX); + ret = ocfs2_refcount_cow(inode, di_bh, cpos, clusters, UINT_MAX); if (ret) mlog_errno(ret); out: @@ -2082,7 +2128,7 @@ static int ocfs2_prepare_inode_for_write(struct file *file, * remove_suid() calls ->setattr without any hint that * we may have already done our cluster locking. Since * ocfs2_setattr() *must* take cluster locks to - * proceeed, this will lead us to recursively lock the + * proceed, this will lead us to recursively lock the * inode. There's also the dinode i_size state which * can be lost via setattr during extending writes (we * set inode->i_size at the end of a write. */ @@ -2187,35 +2233,31 @@ out: return ret; } -static ssize_t ocfs2_file_aio_write(struct kiocb *iocb, - const struct iovec *iov, - unsigned long nr_segs, - loff_t pos) +static ssize_t ocfs2_file_write_iter(struct kiocb *iocb, + struct iov_iter *from) { int ret, direct_io, appending, rw_level, have_alloc_sem = 0; int can_do_direct, has_refcount = 0; ssize_t written = 0; - size_t ocount; /* original count */ - size_t count; /* after file limit checks */ + size_t count = iov_iter_count(from); loff_t old_size, *ppos = &iocb->ki_pos; u32 old_clusters; struct file *file = iocb->ki_filp; - struct inode *inode = file->f_path.dentry->d_inode; + struct inode *inode = file_inode(file); struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); int full_coherency = !(osb->s_mount_opt & OCFS2_MOUNT_COHERENCY_BUFFERED); + int unaligned_dio = 0; trace_ocfs2_file_aio_write(inode, file, file->f_path.dentry, (unsigned long long)OCFS2_I(inode)->ip_blkno, file->f_path.dentry->d_name.len, file->f_path.dentry->d_name.name, - (unsigned int)nr_segs); + (unsigned int)from->nr_segs); /* GRRRRR */ - if (iocb->ki_left == 0) + if (iocb->ki_nbytes == 0) return 0; - vfs_check_frozen(inode->i_sb, SB_FREEZE_WRITE); - appending = file->f_flags & O_APPEND ? 1 : 0; direct_io = file->f_flags & O_DIRECT ? 1 : 0; @@ -2224,9 +2266,8 @@ static ssize_t ocfs2_file_aio_write(struct kiocb *iocb, ocfs2_iocb_clear_sem_locked(iocb); relock: - /* to match setattr's i_mutex -> i_alloc_sem -> rw_lock ordering */ + /* to match setattr's i_mutex -> rw_lock ordering */ if (direct_io) { - down_read(&inode->i_alloc_sem); have_alloc_sem = 1; /* communicate with ocfs2_dio_end_io */ ocfs2_iocb_set_sem_locked(iocb); @@ -2257,7 +2298,7 @@ relock: ret = ocfs2_inode_lock(inode, NULL, 1); if (ret < 0) { mlog_errno(ret); - goto out_sems; + goto out; } ocfs2_inode_unlock(inode, 1); @@ -2265,20 +2306,23 @@ relock: can_do_direct = direct_io; ret = ocfs2_prepare_inode_for_write(file, ppos, - iocb->ki_left, appending, + iocb->ki_nbytes, appending, &can_do_direct, &has_refcount); if (ret < 0) { mlog_errno(ret); goto out; } + if (direct_io && !is_sync_kiocb(iocb)) + unaligned_dio = ocfs2_is_io_unaligned(inode, iocb->ki_nbytes, + *ppos); + /* * We can't complete the direct I/O as requested, fall back to * buffered I/O. */ if (direct_io && !can_do_direct) { ocfs2_rw_unlock(inode, rw_level); - up_read(&inode->i_alloc_sem); have_alloc_sem = 0; rw_level = -1; @@ -2287,6 +2331,16 @@ relock: goto relock; } + if (unaligned_dio) { + /* + * Wait on previous unaligned aio to complete before + * proceeding. + */ + mutex_lock(&OCFS2_I(inode)->ip_unaligned_aio); + /* Mark the iocb as needing an unlock in ocfs2_dio_end_io */ + ocfs2_iocb_set_unaligned_aio(iocb); + } + /* * To later detect whether a journal commit for sync writes is * necessary, we sample i_size, and cluster count here. @@ -2297,28 +2351,23 @@ relock: /* communicate with ocfs2_dio_end_io */ ocfs2_iocb_set_rw_locked(iocb, rw_level); - ret = generic_segment_checks(iov, &nr_segs, &ocount, - VERIFY_READ); - if (ret) - goto out_dio; - - count = ocount; ret = generic_write_checks(file, ppos, &count, S_ISBLK(inode->i_mode)); if (ret) goto out_dio; + iov_iter_truncate(from, count); if (direct_io) { - written = generic_file_direct_write(iocb, iov, &nr_segs, *ppos, - ppos, count, ocount); + written = generic_file_direct_write(iocb, from, *ppos); if (written < 0) { ret = written; goto out_dio; } } else { current->backing_dev_info = file->f_mapping->backing_dev_info; - written = generic_file_buffered_write(iocb, iov, nr_segs, *ppos, - ppos, count, 0); + written = generic_perform_write(file, from, *ppos); + if (likely(written >= 0)) + iocb->ki_pos = *ppos + written; current->backing_dev_info = NULL; } @@ -2328,8 +2377,8 @@ out_dio: if (((file->f_flags & O_DSYNC) && !direct_io) || IS_SYNC(inode) || ((file->f_flags & O_DIRECT) && !direct_io)) { - ret = filemap_fdatawrite_range(file->f_mapping, pos, - pos + count - 1); + ret = filemap_fdatawrite_range(file->f_mapping, *ppos, + *ppos + count - 1); if (ret < 0) written = ret; @@ -2342,15 +2391,14 @@ out_dio: } if (!ret) - ret = filemap_fdatawait_range(file->f_mapping, pos, - pos + count - 1); + ret = filemap_fdatawait_range(file->f_mapping, *ppos, + *ppos + count - 1); } /* * deep in g_f_a_w_n()->ocfs2_direct_IO we pass in a ocfs2_dio_end_io * function pointer which is called when o_direct io completes so that - * it can unlock our rw lock. (it's the clustered equivalent of - * i_alloc_sem; protects truncate from racing with pending ios). + * it can unlock our rw lock. * Unfortunately there are error cases which call end_io and others * that don't. so we don't have to unlock the rw_lock if either an * async dio is going to do it in the future or an end_io after an @@ -2359,6 +2407,12 @@ out_dio: if ((ret == -EIOCBQUEUED) || (!ocfs2_iocb_is_rw_locked(iocb))) { rw_level = -1; have_alloc_sem = 0; + unaligned_dio = 0; + } + + if (unaligned_dio) { + ocfs2_iocb_clear_unaligned_aio(iocb); + mutex_unlock(&OCFS2_I(inode)->ip_unaligned_aio); } out: @@ -2366,10 +2420,8 @@ out: ocfs2_rw_unlock(inode, rw_level); out_sems: - if (have_alloc_sem) { - up_read(&inode->i_alloc_sem); + if (have_alloc_sem) ocfs2_iocb_clear_sem_locked(iocb); - } mutex_unlock(&inode->i_mutex); @@ -2378,89 +2430,6 @@ out_sems: return ret; } -static int ocfs2_splice_to_file(struct pipe_inode_info *pipe, - struct file *out, - struct splice_desc *sd) -{ - int ret; - - ret = ocfs2_prepare_inode_for_write(out, &sd->pos, - sd->total_len, 0, NULL, NULL); - if (ret < 0) { - mlog_errno(ret); - return ret; - } - - return splice_from_pipe_feed(pipe, sd, pipe_to_file); -} - -static ssize_t ocfs2_file_splice_write(struct pipe_inode_info *pipe, - struct file *out, - loff_t *ppos, - size_t len, - unsigned int flags) -{ - int ret; - struct address_space *mapping = out->f_mapping; - struct inode *inode = mapping->host; - struct splice_desc sd = { - .total_len = len, - .flags = flags, - .pos = *ppos, - .u.file = out, - }; - - - trace_ocfs2_file_splice_write(inode, out, out->f_path.dentry, - (unsigned long long)OCFS2_I(inode)->ip_blkno, - out->f_path.dentry->d_name.len, - out->f_path.dentry->d_name.name, len); - - if (pipe->inode) - mutex_lock_nested(&pipe->inode->i_mutex, I_MUTEX_PARENT); - - splice_from_pipe_begin(&sd); - do { - ret = splice_from_pipe_next(pipe, &sd); - if (ret <= 0) - break; - - mutex_lock_nested(&inode->i_mutex, I_MUTEX_CHILD); - ret = ocfs2_rw_lock(inode, 1); - if (ret < 0) - mlog_errno(ret); - else { - ret = ocfs2_splice_to_file(pipe, out, &sd); - ocfs2_rw_unlock(inode, 1); - } - mutex_unlock(&inode->i_mutex); - } while (ret > 0); - splice_from_pipe_end(pipe, &sd); - - if (pipe->inode) - mutex_unlock(&pipe->inode->i_mutex); - - if (sd.num_spliced) - ret = sd.num_spliced; - - if (ret > 0) { - unsigned long nr_pages; - int err; - - nr_pages = (ret + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT; - - err = generic_write_sync(out, *ppos, ret); - if (err) - ret = err; - else - *ppos += ret; - - balance_dirty_pages_ratelimited_nr(mapping, nr_pages); - } - - return ret; -} - static ssize_t ocfs2_file_splice_read(struct file *in, loff_t *ppos, struct pipe_inode_info *pipe, @@ -2468,7 +2437,7 @@ static ssize_t ocfs2_file_splice_read(struct file *in, unsigned int flags) { int ret = 0, lock_level = 0; - struct inode *inode = in->f_path.dentry->d_inode; + struct inode *inode = file_inode(in); trace_ocfs2_file_splice_read(inode, in, in->f_path.dentry, (unsigned long long)OCFS2_I(inode)->ip_blkno, @@ -2476,9 +2445,9 @@ static ssize_t ocfs2_file_splice_read(struct file *in, in->f_path.dentry->d_name.name, len); /* - * See the comment in ocfs2_file_aio_read() + * See the comment in ocfs2_file_read_iter() */ - ret = ocfs2_inode_lock_atime(inode, in->f_vfsmnt, &lock_level); + ret = ocfs2_inode_lock_atime(inode, in->f_path.mnt, &lock_level); if (ret < 0) { mlog_errno(ret); goto bail; @@ -2491,19 +2460,18 @@ bail: return ret; } -static ssize_t ocfs2_file_aio_read(struct kiocb *iocb, - const struct iovec *iov, - unsigned long nr_segs, - loff_t pos) +static ssize_t ocfs2_file_read_iter(struct kiocb *iocb, + struct iov_iter *to) { int ret = 0, rw_level = -1, have_alloc_sem = 0, lock_level = 0; struct file *filp = iocb->ki_filp; - struct inode *inode = filp->f_path.dentry->d_inode; + struct inode *inode = file_inode(filp); trace_ocfs2_file_aio_read(inode, filp, filp->f_path.dentry, (unsigned long long)OCFS2_I(inode)->ip_blkno, filp->f_path.dentry->d_name.len, - filp->f_path.dentry->d_name.name, nr_segs); + filp->f_path.dentry->d_name.name, + to->nr_segs); /* GRRRRR */ if (!inode) { @@ -2519,7 +2487,6 @@ static ssize_t ocfs2_file_aio_read(struct kiocb *iocb, * need locks to protect pending reads from racing with truncate. */ if (filp->f_flags & O_DIRECT) { - down_read(&inode->i_alloc_sem); have_alloc_sem = 1; ocfs2_iocb_set_sem_locked(iocb); @@ -2542,36 +2509,85 @@ static ssize_t ocfs2_file_aio_read(struct kiocb *iocb, * like i_size. This allows the checks down below * generic_file_aio_read() a chance of actually working. */ - ret = ocfs2_inode_lock_atime(inode, filp->f_vfsmnt, &lock_level); + ret = ocfs2_inode_lock_atime(inode, filp->f_path.mnt, &lock_level); if (ret < 0) { mlog_errno(ret); goto bail; } ocfs2_inode_unlock(inode, lock_level); - ret = generic_file_aio_read(iocb, iov, nr_segs, iocb->ki_pos); + ret = generic_file_read_iter(iocb, to); trace_generic_file_aio_read_ret(ret); /* buffered aio wouldn't have proper lock coverage today */ BUG_ON(ret == -EIOCBQUEUED && !(filp->f_flags & O_DIRECT)); - /* see ocfs2_file_aio_write */ + /* see ocfs2_file_write_iter */ if (ret == -EIOCBQUEUED || !ocfs2_iocb_is_rw_locked(iocb)) { rw_level = -1; have_alloc_sem = 0; } bail: - if (have_alloc_sem) { - up_read(&inode->i_alloc_sem); + if (have_alloc_sem) ocfs2_iocb_clear_sem_locked(iocb); - } + if (rw_level != -1) ocfs2_rw_unlock(inode, rw_level); return ret; } +/* Refer generic_file_llseek_unlocked() */ +static loff_t ocfs2_file_llseek(struct file *file, loff_t offset, int whence) +{ + struct inode *inode = file->f_mapping->host; + int ret = 0; + + mutex_lock(&inode->i_mutex); + + switch (whence) { + case SEEK_SET: + break; + case SEEK_END: + /* SEEK_END requires the OCFS2 inode lock for the file + * because it references the file's size. + */ + ret = ocfs2_inode_lock(inode, NULL, 0); + if (ret < 0) { + mlog_errno(ret); + goto out; + } + offset += i_size_read(inode); + ocfs2_inode_unlock(inode, 0); + break; + case SEEK_CUR: + if (offset == 0) { + offset = file->f_pos; + goto out; + } + offset += file->f_pos; + break; + case SEEK_DATA: + case SEEK_HOLE: + ret = ocfs2_seek_data_hole_offset(file, &offset, whence); + if (ret) + goto out; + break; + default: + ret = -EINVAL; + goto out; + } + + offset = vfs_setpos(file, offset, inode->i_sb->s_maxbytes); + +out: + mutex_unlock(&inode->i_mutex); + if (ret) + return ret; + return offset; +} + const struct inode_operations ocfs2_file_iops = { .setattr = ocfs2_setattr, .getattr = ocfs2_getattr, @@ -2581,12 +2597,16 @@ const struct inode_operations ocfs2_file_iops = { .listxattr = ocfs2_listxattr, .removexattr = generic_removexattr, .fiemap = ocfs2_fiemap, + .get_acl = ocfs2_iop_get_acl, + .set_acl = ocfs2_iop_set_acl, }; const struct inode_operations ocfs2_special_file_iops = { .setattr = ocfs2_setattr, .getattr = ocfs2_getattr, .permission = ocfs2_permission, + .get_acl = ocfs2_iop_get_acl, + .set_acl = ocfs2_iop_set_acl, }; /* @@ -2594,15 +2614,15 @@ const struct inode_operations ocfs2_special_file_iops = { * ocfs2_fops_no_plocks and ocfs2_dops_no_plocks! */ const struct file_operations ocfs2_fops = { - .llseek = generic_file_llseek, - .read = do_sync_read, - .write = do_sync_write, + .llseek = ocfs2_file_llseek, + .read = new_sync_read, + .write = new_sync_write, .mmap = ocfs2_mmap, .fsync = ocfs2_sync_file, .release = ocfs2_file_release, .open = ocfs2_file_open, - .aio_read = ocfs2_file_aio_read, - .aio_write = ocfs2_file_aio_write, + .read_iter = ocfs2_file_read_iter, + .write_iter = ocfs2_file_write_iter, .unlocked_ioctl = ocfs2_ioctl, #ifdef CONFIG_COMPAT .compat_ioctl = ocfs2_compat_ioctl, @@ -2610,14 +2630,14 @@ const struct file_operations ocfs2_fops = { .lock = ocfs2_lock, .flock = ocfs2_flock, .splice_read = ocfs2_file_splice_read, - .splice_write = ocfs2_file_splice_write, + .splice_write = iter_file_splice_write, .fallocate = ocfs2_fallocate, }; const struct file_operations ocfs2_dops = { .llseek = generic_file_llseek, .read = generic_read_dir, - .readdir = ocfs2_readdir, + .iterate = ocfs2_readdir, .fsync = ocfs2_sync_file, .release = ocfs2_dir_release, .open = ocfs2_dir_open, @@ -2642,28 +2662,29 @@ const struct file_operations ocfs2_dops = { * the cluster. */ const struct file_operations ocfs2_fops_no_plocks = { - .llseek = generic_file_llseek, - .read = do_sync_read, - .write = do_sync_write, + .llseek = ocfs2_file_llseek, + .read = new_sync_read, + .write = new_sync_write, .mmap = ocfs2_mmap, .fsync = ocfs2_sync_file, .release = ocfs2_file_release, .open = ocfs2_file_open, - .aio_read = ocfs2_file_aio_read, - .aio_write = ocfs2_file_aio_write, + .read_iter = ocfs2_file_read_iter, + .write_iter = ocfs2_file_write_iter, .unlocked_ioctl = ocfs2_ioctl, #ifdef CONFIG_COMPAT .compat_ioctl = ocfs2_compat_ioctl, #endif .flock = ocfs2_flock, .splice_read = ocfs2_file_splice_read, - .splice_write = ocfs2_file_splice_write, + .splice_write = iter_file_splice_write, + .fallocate = ocfs2_fallocate, }; const struct file_operations ocfs2_dops_no_plocks = { .llseek = generic_file_llseek, .read = generic_read_dir, - .readdir = ocfs2_readdir, + .iterate = ocfs2_readdir, .fsync = ocfs2_sync_file, .release = ocfs2_dir_release, .open = ocfs2_dir_open, diff --git a/fs/ocfs2/file.h b/fs/ocfs2/file.h index f5afbbef670..97bf761c9e7 100644 --- a/fs/ocfs2/file.h +++ b/fs/ocfs2/file.h @@ -61,7 +61,7 @@ int ocfs2_zero_extend(struct inode *inode, struct buffer_head *di_bh, int ocfs2_setattr(struct dentry *dentry, struct iattr *attr); int ocfs2_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat); -int ocfs2_permission(struct inode *inode, int mask, unsigned int flags); +int ocfs2_permission(struct inode *inode, int mask); int ocfs2_should_update_atime(struct inode *inode, struct vfsmount *vfsmnt); diff --git a/fs/ocfs2/inode.c b/fs/ocfs2/inode.c index b4c8bb6b8d2..437de7f768c 100644 --- a/fs/ocfs2/inode.c +++ b/fs/ocfs2/inode.c @@ -130,6 +130,7 @@ struct inode *ocfs2_iget(struct ocfs2_super *osb, u64 blkno, unsigned flags, struct inode *inode = NULL; struct super_block *sb = osb->sb; struct ocfs2_find_inode_args args; + journal_t *journal = OCFS2_SB(sb)->journal->j_journal; trace_ocfs2_iget_begin((unsigned long long)blkno, flags, sysfile_type); @@ -169,6 +170,32 @@ struct inode *ocfs2_iget(struct ocfs2_super *osb, u64 blkno, unsigned flags, goto bail; } + /* + * Set transaction id's of transactions that have to be committed + * to finish f[data]sync. We set them to currently running transaction + * as we cannot be sure that the inode or some of its metadata isn't + * part of the transaction - the inode could have been reclaimed and + * now it is reread from disk. + */ + if (journal) { + transaction_t *transaction; + tid_t tid; + struct ocfs2_inode_info *oi = OCFS2_I(inode); + + read_lock(&journal->j_state_lock); + if (journal->j_running_transaction) + transaction = journal->j_running_transaction; + else + transaction = journal->j_committing_transaction; + if (transaction) + tid = transaction->t_tid; + else + tid = journal->j_commit_sequence; + read_unlock(&journal->j_state_lock); + oi->i_sync_tid = tid; + oi->i_datasync_tid = tid; + } + bail: if (!IS_ERR(inode)) { trace_ocfs2_iget_end(inode, @@ -269,15 +296,17 @@ void ocfs2_populate_inode(struct inode *inode, struct ocfs2_dinode *fe, inode->i_generation = le32_to_cpu(fe->i_generation); inode->i_rdev = huge_decode_dev(le64_to_cpu(fe->id1.dev1.i_rdev)); inode->i_mode = le16_to_cpu(fe->i_mode); - inode->i_uid = le32_to_cpu(fe->i_uid); - inode->i_gid = le32_to_cpu(fe->i_gid); + i_uid_write(inode, le32_to_cpu(fe->i_uid)); + i_gid_write(inode, le32_to_cpu(fe->i_gid)); /* Fast symlinks will have i_size but no allocated clusters. */ - if (S_ISLNK(inode->i_mode) && !fe->i_clusters) + if (S_ISLNK(inode->i_mode) && !fe->i_clusters) { inode->i_blocks = 0; - else + inode->i_mapping->a_ops = &ocfs2_fast_symlink_aops; + } else { inode->i_blocks = ocfs2_inode_sector_count(inode); - inode->i_mapping->a_ops = &ocfs2_aops; + inode->i_mapping->a_ops = &ocfs2_aops; + } inode->i_atime.tv_sec = le64_to_cpu(fe->i_atime); inode->i_atime.tv_nsec = le32_to_cpu(fe->i_atime_nsec); inode->i_mtime.tv_sec = le64_to_cpu(fe->i_mtime); @@ -291,7 +320,7 @@ void ocfs2_populate_inode(struct inode *inode, struct ocfs2_dinode *fe, (unsigned long long)OCFS2_I(inode)->ip_blkno, (unsigned long long)le64_to_cpu(fe->i_blkno)); - inode->i_nlink = ocfs2_read_links_count(fe); + set_nlink(inode, ocfs2_read_links_count(fe)); trace_ocfs2_populate_inode(OCFS2_I(inode)->ip_blkno, le32_to_cpu(fe->i_flags)); @@ -331,10 +360,7 @@ void ocfs2_populate_inode(struct inode *inode, struct ocfs2_dinode *fe, OCFS2_I(inode)->ip_dir_lock_gen = 1; break; case S_IFLNK: - if (ocfs2_inode_is_fast_symlink(inode)) - inode->i_op = &ocfs2_fast_symlink_inode_operations; - else - inode->i_op = &ocfs2_symlink_inode_operations; + inode->i_op = &ocfs2_symlink_inode_operations; i_size_write(inode, le64_to_cpu(fe->i_size)); break; default: @@ -387,19 +413,9 @@ static int ocfs2_read_locked_inode(struct inode *inode, u32 generation = 0; status = -EINVAL; - if (inode == NULL || inode->i_sb == NULL) { - mlog(ML_ERROR, "bad inode\n"); - return status; - } sb = inode->i_sb; osb = OCFS2_SB(sb); - if (!args) { - mlog(ML_ERROR, "bad inode args\n"); - make_bad_inode(inode); - return status; - } - /* * To improve performance of cold-cache inode stats, we take * the cluster lock here if possible. @@ -815,11 +831,13 @@ static int ocfs2_inode_is_valid_to_delete(struct inode *inode) goto bail; } - /* If we're coming from downconvert_thread we can't go into our own - * voting [hello, deadlock city!], so unforuntately we just - * have to skip deleting this guy. That's OK though because - * the node who's doing the actual deleting should handle it - * anyway. */ + /* + * If we're coming from downconvert_thread we can't go into our own + * voting [hello, deadlock city!] so we cannot delete the inode. But + * since we dropped last inode ref when downconverting dentry lock, + * we cannot have the file open and thus the node doing unlink will + * take care of deleting the inode. + */ if (current == osb->dc_task) goto bail; @@ -833,12 +851,6 @@ static int ocfs2_inode_is_valid_to_delete(struct inode *inode) goto bail_unlock; } - /* If we have allowd wipe of this inode for another node, it - * will be marked here so we can safely skip it. Recovery will - * cleanup any inodes we might inadvertently skip here. */ - if (oi->ip_flags & OCFS2_INODE_SKIP_DELETE) - goto bail_unlock; - ret = 1; bail_unlock: spin_unlock(&oi->ip_lock); @@ -951,8 +963,8 @@ static void ocfs2_cleanup_delete_inode(struct inode *inode, trace_ocfs2_cleanup_delete_inode( (unsigned long long)OCFS2_I(inode)->ip_blkno, sync_data); if (sync_data) - write_inode_now(inode, 1); - truncate_inode_pages(&inode->i_data, 0); + filemap_write_and_wait(inode->i_mapping); + truncate_inode_pages_final(&inode->i_data); } static void ocfs2_delete_inode(struct inode *inode) @@ -971,8 +983,6 @@ static void ocfs2_delete_inode(struct inode *inode) if (is_bad_inode(inode) || !OCFS2_I(inode)->ip_blkno) goto bail; - dquot_initialize(inode); - if (!ocfs2_inode_is_valid_to_delete(inode)) { /* It's probably not necessary to truncate_inode_pages * here but we do it for safety anyway (it will most @@ -981,6 +991,8 @@ static void ocfs2_delete_inode(struct inode *inode) goto bail; } + dquot_initialize(inode); + /* We want to block signals in delete_inode as the lock and * messaging paths may return us -ERESTARTSYS. Which would * cause us to exit early, resulting in inodes being orphaned @@ -1068,8 +1080,9 @@ static void ocfs2_clear_inode(struct inode *inode) { int status; struct ocfs2_inode_info *oi = OCFS2_I(inode); + struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); - end_writeback(inode); + clear_inode(inode); trace_ocfs2_clear_inode((unsigned long long)oi->ip_blkno, inode->i_nlink); @@ -1084,9 +1097,9 @@ static void ocfs2_clear_inode(struct inode *inode) /* Do these before all the other work so that we don't bounce * the downconvert thread while waiting to destroy the locks. */ - ocfs2_mark_lockres_freeing(&oi->ip_rw_lockres); - ocfs2_mark_lockres_freeing(&oi->ip_inode_lockres); - ocfs2_mark_lockres_freeing(&oi->ip_open_lockres); + ocfs2_mark_lockres_freeing(osb, &oi->ip_rw_lockres); + ocfs2_mark_lockres_freeing(osb, &oi->ip_inode_lockres); + ocfs2_mark_lockres_freeing(osb, &oi->ip_open_lockres); ocfs2_resv_discard(&OCFS2_SB(inode->i_sb)->osb_la_resmap, &oi->ip_la_data_resv); @@ -1168,7 +1181,7 @@ void ocfs2_evict_inode(struct inode *inode) (OCFS2_I(inode)->ip_flags & OCFS2_INODE_MAYBE_ORPHANED)) { ocfs2_delete_inode(inode); } else { - truncate_inode_pages(&inode->i_data, 0); + truncate_inode_pages_final(&inode->i_data); } ocfs2_clear_inode(inode); } @@ -1260,8 +1273,8 @@ int ocfs2_mark_inode_dirty(handle_t *handle, fe->i_size = cpu_to_le64(i_size_read(inode)); ocfs2_set_links_count(fe, inode->i_nlink); - fe->i_uid = cpu_to_le32(inode->i_uid); - fe->i_gid = cpu_to_le32(inode->i_gid); + fe->i_uid = cpu_to_le32(i_uid_read(inode)); + fe->i_gid = cpu_to_le32(i_gid_read(inode)); fe->i_mode = cpu_to_le16(inode->i_mode); fe->i_atime = cpu_to_le64(inode->i_atime.tv_sec); fe->i_atime_nsec = cpu_to_le32(inode->i_atime.tv_nsec); @@ -1271,6 +1284,7 @@ int ocfs2_mark_inode_dirty(handle_t *handle, fe->i_mtime_nsec = cpu_to_le32(inode->i_mtime.tv_nsec); ocfs2_journal_dirty(handle, bh); + ocfs2_update_inode_fsync_trans(handle, inode, 1); leave: return status; } @@ -1290,9 +1304,9 @@ void ocfs2_refresh_inode(struct inode *inode, OCFS2_I(inode)->ip_dyn_features = le16_to_cpu(fe->i_dyn_features); ocfs2_set_inode_flags(inode); i_size_write(inode, le64_to_cpu(fe->i_size)); - inode->i_nlink = ocfs2_read_links_count(fe); - inode->i_uid = le32_to_cpu(fe->i_uid); - inode->i_gid = le32_to_cpu(fe->i_gid); + set_nlink(inode, ocfs2_read_links_count(fe)); + i_uid_write(inode, le32_to_cpu(fe->i_uid)); + i_gid_write(inode, le32_to_cpu(fe->i_gid)); inode->i_mode = le16_to_cpu(fe->i_mode); if (S_ISLNK(inode->i_mode) && le32_to_cpu(fe->i_clusters) == 0) inode->i_blocks = 0; diff --git a/fs/ocfs2/inode.h b/fs/ocfs2/inode.h index 1c508b149b3..a6c991c0fc9 100644 --- a/fs/ocfs2/inode.h +++ b/fs/ocfs2/inode.h @@ -43,6 +43,9 @@ struct ocfs2_inode_info /* protects extended attribute changes on this inode */ struct rw_semaphore ip_xattr_sem; + /* Number of outstanding AIO's which are not page aligned */ + struct mutex ip_unaligned_aio; + /* These fields are protected by ip_lock */ spinlock_t ip_lock; u32 ip_open_count; @@ -70,6 +73,13 @@ struct ocfs2_inode_info u32 ip_dir_lock_gen; struct ocfs2_alloc_reservation ip_la_data_resv; + + /* + * Transactions that contain inode's metadata needed to complete + * fsync and fdatasync, respectively. + */ + tid_t i_sync_tid; + tid_t i_datasync_tid; }; /* @@ -81,8 +91,6 @@ struct ocfs2_inode_info #define OCFS2_INODE_BITMAP 0x00000004 /* This inode has been wiped from disk */ #define OCFS2_INODE_DELETED 0x00000008 -/* Another node is deleting, so our delete is a nop */ -#define OCFS2_INODE_SKIP_DELETE 0x00000010 /* Has the inode been orphaned on another node? * * This hints to ocfs2_drop_inode that it should clear i_nlink before @@ -97,11 +105,11 @@ struct ocfs2_inode_info * rely on ocfs2_delete_inode to sort things out under the proper * cluster locks. */ -#define OCFS2_INODE_MAYBE_ORPHANED 0x00000020 +#define OCFS2_INODE_MAYBE_ORPHANED 0x00000010 /* Does someone have the file open O_DIRECT */ -#define OCFS2_INODE_OPEN_DIRECT 0x00000040 +#define OCFS2_INODE_OPEN_DIRECT 0x00000020 /* Tell the inode wipe code it's not in orphan dir */ -#define OCFS2_INODE_SKIP_ORPHAN_DIR 0x00000080 +#define OCFS2_INODE_SKIP_ORPHAN_DIR 0x00000040 static inline struct ocfs2_inode_info *OCFS2_I(struct inode *inode) { @@ -144,8 +152,6 @@ void ocfs2_refresh_inode(struct inode *inode, int ocfs2_mark_inode_dirty(handle_t *handle, struct inode *inode, struct buffer_head *bh); -int ocfs2_aio_read(struct file *file, struct kiocb *req, struct iocb *iocb); -int ocfs2_aio_write(struct file *file, struct kiocb *req, struct iocb *iocb); struct buffer_head *ocfs2_bread(struct inode *inode, int block, int *err, int reada); diff --git a/fs/ocfs2/ioctl.c b/fs/ocfs2/ioctl.c index 8f13c5989ea..6f66b3751ac 100644 --- a/fs/ocfs2/ioctl.c +++ b/fs/ocfs2/ioctl.c @@ -7,6 +7,7 @@ #include <linux/fs.h> #include <linux/mount.h> +#include <linux/blkdev.h> #include <linux/compat.h> #include <cluster/masklog.h> @@ -22,8 +23,11 @@ #include "ioctl.h" #include "resize.h" #include "refcounttree.h" - -#include <linux/ext2_fs.h> +#include "sysfile.h" +#include "dir.h" +#include "buffer_head_io.h" +#include "suballoc.h" +#include "move_extents.h" #define o2info_from_user(a, b) \ copy_from_user(&(a), (b), sizeof(a)) @@ -35,31 +39,27 @@ * be -EFAULT. The error will be returned from the ioctl(2) call. It's * just a best-effort to tell userspace that this request caused the error. */ -static inline void __o2info_set_request_error(struct ocfs2_info_request *kreq, +static inline void o2info_set_request_error(struct ocfs2_info_request *kreq, struct ocfs2_info_request __user *req) { kreq->ir_flags |= OCFS2_INFO_FL_ERROR; (void)put_user(kreq->ir_flags, (__u32 __user *)&(req->ir_flags)); } -#define o2info_set_request_error(a, b) \ - __o2info_set_request_error((struct ocfs2_info_request *)&(a), b) - -static inline void __o2info_set_request_filled(struct ocfs2_info_request *req) +static inline void o2info_set_request_filled(struct ocfs2_info_request *req) { req->ir_flags |= OCFS2_INFO_FL_FILLED; } -#define o2info_set_request_filled(a) \ - __o2info_set_request_filled((struct ocfs2_info_request *)&(a)) - -static inline void __o2info_clear_request_filled(struct ocfs2_info_request *req) +static inline void o2info_clear_request_filled(struct ocfs2_info_request *req) { req->ir_flags &= ~OCFS2_INFO_FL_FILLED; } -#define o2info_clear_request_filled(a) \ - __o2info_clear_request_filled((struct ocfs2_info_request *)&(a)) +static inline int o2info_coherent(struct ocfs2_info_request *req) +{ + return (!(req->ir_flags & OCFS2_INFO_FL_NON_COHERENT)); +} static int ocfs2_get_inode_attr(struct inode *inode, unsigned *flags) { @@ -102,13 +102,6 @@ static int ocfs2_set_inode_attr(struct inode *inode, unsigned flags, if (!S_ISDIR(inode->i_mode)) flags &= ~OCFS2_DIRSYNC_FL; - handle = ocfs2_start_trans(osb, OCFS2_INODE_UPDATE_CREDITS); - if (IS_ERR(handle)) { - status = PTR_ERR(handle); - mlog_errno(status); - goto bail_unlock; - } - oldflags = ocfs2_inode->ip_attr; flags = flags & mask; flags |= oldflags & ~mask; @@ -124,6 +117,13 @@ static int ocfs2_set_inode_attr(struct inode *inode, unsigned flags, goto bail_unlock; } + handle = ocfs2_start_trans(osb, OCFS2_INODE_UPDATE_CREDITS); + if (IS_ERR(handle)) { + status = PTR_ERR(handle); + mlog_errno(status); + goto bail_unlock; + } + ocfs2_inode->ip_attr = flags; ocfs2_set_inode_flags(inode); @@ -132,6 +132,7 @@ static int ocfs2_set_inode_attr(struct inode *inode, unsigned flags, mlog_errno(status); ocfs2_commit_trans(osb, handle); + bail_unlock: ocfs2_inode_unlock(inode, 1); bail: @@ -142,8 +143,8 @@ bail: return status; } -int ocfs2_info_handle_blocksize(struct inode *inode, - struct ocfs2_info_request __user *req) +static int ocfs2_info_handle_blocksize(struct inode *inode, + struct ocfs2_info_request __user *req) { int status = -EFAULT; struct ocfs2_info_blocksize oib; @@ -153,7 +154,7 @@ int ocfs2_info_handle_blocksize(struct inode *inode, oib.ib_blocksize = inode->i_sb->s_blocksize; - o2info_set_request_filled(oib); + o2info_set_request_filled(&oib.ib_req); if (o2info_to_user(oib, req)) goto bail; @@ -161,13 +162,13 @@ int ocfs2_info_handle_blocksize(struct inode *inode, status = 0; bail: if (status) - o2info_set_request_error(oib, req); + o2info_set_request_error(&oib.ib_req, req); return status; } -int ocfs2_info_handle_clustersize(struct inode *inode, - struct ocfs2_info_request __user *req) +static int ocfs2_info_handle_clustersize(struct inode *inode, + struct ocfs2_info_request __user *req) { int status = -EFAULT; struct ocfs2_info_clustersize oic; @@ -178,7 +179,7 @@ int ocfs2_info_handle_clustersize(struct inode *inode, oic.ic_clustersize = osb->s_clustersize; - o2info_set_request_filled(oic); + o2info_set_request_filled(&oic.ic_req); if (o2info_to_user(oic, req)) goto bail; @@ -186,13 +187,13 @@ int ocfs2_info_handle_clustersize(struct inode *inode, status = 0; bail: if (status) - o2info_set_request_error(oic, req); + o2info_set_request_error(&oic.ic_req, req); return status; } -int ocfs2_info_handle_maxslots(struct inode *inode, - struct ocfs2_info_request __user *req) +static int ocfs2_info_handle_maxslots(struct inode *inode, + struct ocfs2_info_request __user *req) { int status = -EFAULT; struct ocfs2_info_maxslots oim; @@ -203,7 +204,7 @@ int ocfs2_info_handle_maxslots(struct inode *inode, oim.im_max_slots = osb->max_slots; - o2info_set_request_filled(oim); + o2info_set_request_filled(&oim.im_req); if (o2info_to_user(oim, req)) goto bail; @@ -211,13 +212,13 @@ int ocfs2_info_handle_maxslots(struct inode *inode, status = 0; bail: if (status) - o2info_set_request_error(oim, req); + o2info_set_request_error(&oim.im_req, req); return status; } -int ocfs2_info_handle_label(struct inode *inode, - struct ocfs2_info_request __user *req) +static int ocfs2_info_handle_label(struct inode *inode, + struct ocfs2_info_request __user *req) { int status = -EFAULT; struct ocfs2_info_label oil; @@ -228,7 +229,7 @@ int ocfs2_info_handle_label(struct inode *inode, memcpy(oil.il_label, osb->vol_label, OCFS2_MAX_VOL_LABEL_LEN); - o2info_set_request_filled(oil); + o2info_set_request_filled(&oil.il_req); if (o2info_to_user(oil, req)) goto bail; @@ -236,13 +237,13 @@ int ocfs2_info_handle_label(struct inode *inode, status = 0; bail: if (status) - o2info_set_request_error(oil, req); + o2info_set_request_error(&oil.il_req, req); return status; } -int ocfs2_info_handle_uuid(struct inode *inode, - struct ocfs2_info_request __user *req) +static int ocfs2_info_handle_uuid(struct inode *inode, + struct ocfs2_info_request __user *req) { int status = -EFAULT; struct ocfs2_info_uuid oiu; @@ -253,7 +254,7 @@ int ocfs2_info_handle_uuid(struct inode *inode, memcpy(oiu.iu_uuid_str, osb->uuid_str, OCFS2_TEXT_UUID_LEN + 1); - o2info_set_request_filled(oiu); + o2info_set_request_filled(&oiu.iu_req); if (o2info_to_user(oiu, req)) goto bail; @@ -261,13 +262,13 @@ int ocfs2_info_handle_uuid(struct inode *inode, status = 0; bail: if (status) - o2info_set_request_error(oiu, req); + o2info_set_request_error(&oiu.iu_req, req); return status; } -int ocfs2_info_handle_fs_features(struct inode *inode, - struct ocfs2_info_request __user *req) +static int ocfs2_info_handle_fs_features(struct inode *inode, + struct ocfs2_info_request __user *req) { int status = -EFAULT; struct ocfs2_info_fs_features oif; @@ -280,7 +281,7 @@ int ocfs2_info_handle_fs_features(struct inode *inode, oif.if_incompat_features = osb->s_feature_incompat; oif.if_ro_compat_features = osb->s_feature_ro_compat; - o2info_set_request_filled(oif); + o2info_set_request_filled(&oif.if_req); if (o2info_to_user(oif, req)) goto bail; @@ -288,13 +289,13 @@ int ocfs2_info_handle_fs_features(struct inode *inode, status = 0; bail: if (status) - o2info_set_request_error(oif, req); + o2info_set_request_error(&oif.if_req, req); return status; } -int ocfs2_info_handle_journal_size(struct inode *inode, - struct ocfs2_info_request __user *req) +static int ocfs2_info_handle_journal_size(struct inode *inode, + struct ocfs2_info_request __user *req) { int status = -EFAULT; struct ocfs2_info_journal_size oij; @@ -303,9 +304,9 @@ int ocfs2_info_handle_journal_size(struct inode *inode, if (o2info_from_user(oij, req)) goto bail; - oij.ij_journal_size = osb->journal->j_inode->i_size; + oij.ij_journal_size = i_size_read(osb->journal->j_inode); - o2info_set_request_filled(oij); + o2info_set_request_filled(&oij.ij_req); if (o2info_to_user(oij, req)) goto bail; @@ -313,13 +314,418 @@ int ocfs2_info_handle_journal_size(struct inode *inode, status = 0; bail: if (status) - o2info_set_request_error(oij, req); + o2info_set_request_error(&oij.ij_req, req); + + return status; +} + +static int ocfs2_info_scan_inode_alloc(struct ocfs2_super *osb, + struct inode *inode_alloc, u64 blkno, + struct ocfs2_info_freeinode *fi, + u32 slot) +{ + int status = 0, unlock = 0; + + struct buffer_head *bh = NULL; + struct ocfs2_dinode *dinode_alloc = NULL; + + if (inode_alloc) + mutex_lock(&inode_alloc->i_mutex); + + if (o2info_coherent(&fi->ifi_req)) { + status = ocfs2_inode_lock(inode_alloc, &bh, 0); + if (status < 0) { + mlog_errno(status); + goto bail; + } + unlock = 1; + } else { + status = ocfs2_read_blocks_sync(osb, blkno, 1, &bh); + if (status < 0) { + mlog_errno(status); + goto bail; + } + } + + dinode_alloc = (struct ocfs2_dinode *)bh->b_data; + + fi->ifi_stat[slot].lfi_total = + le32_to_cpu(dinode_alloc->id1.bitmap1.i_total); + fi->ifi_stat[slot].lfi_free = + le32_to_cpu(dinode_alloc->id1.bitmap1.i_total) - + le32_to_cpu(dinode_alloc->id1.bitmap1.i_used); + +bail: + if (unlock) + ocfs2_inode_unlock(inode_alloc, 0); + + if (inode_alloc) + mutex_unlock(&inode_alloc->i_mutex); + + brelse(bh); + + return status; +} + +static int ocfs2_info_handle_freeinode(struct inode *inode, + struct ocfs2_info_request __user *req) +{ + u32 i; + u64 blkno = -1; + char namebuf[40]; + int status = -EFAULT, type = INODE_ALLOC_SYSTEM_INODE; + struct ocfs2_info_freeinode *oifi = NULL; + struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); + struct inode *inode_alloc = NULL; + + oifi = kzalloc(sizeof(struct ocfs2_info_freeinode), GFP_KERNEL); + if (!oifi) { + status = -ENOMEM; + mlog_errno(status); + goto out_err; + } + + if (o2info_from_user(*oifi, req)) + goto bail; + + oifi->ifi_slotnum = osb->max_slots; + + for (i = 0; i < oifi->ifi_slotnum; i++) { + if (o2info_coherent(&oifi->ifi_req)) { + inode_alloc = ocfs2_get_system_file_inode(osb, type, i); + if (!inode_alloc) { + mlog(ML_ERROR, "unable to get alloc inode in " + "slot %u\n", i); + status = -EIO; + goto bail; + } + } else { + ocfs2_sprintf_system_inode_name(namebuf, + sizeof(namebuf), + type, i); + status = ocfs2_lookup_ino_from_name(osb->sys_root_inode, + namebuf, + strlen(namebuf), + &blkno); + if (status < 0) { + status = -ENOENT; + goto bail; + } + } + + status = ocfs2_info_scan_inode_alloc(osb, inode_alloc, blkno, oifi, i); + + iput(inode_alloc); + inode_alloc = NULL; + + if (status < 0) + goto bail; + } + + o2info_set_request_filled(&oifi->ifi_req); + + if (o2info_to_user(*oifi, req)) + goto bail; + + status = 0; +bail: + if (status) + o2info_set_request_error(&oifi->ifi_req, req); + + kfree(oifi); +out_err: + return status; +} + +static void o2ffg_update_histogram(struct ocfs2_info_free_chunk_list *hist, + unsigned int chunksize) +{ + int index; + + index = __ilog2_u32(chunksize); + if (index >= OCFS2_INFO_MAX_HIST) + index = OCFS2_INFO_MAX_HIST - 1; + + hist->fc_chunks[index]++; + hist->fc_clusters[index] += chunksize; +} + +static void o2ffg_update_stats(struct ocfs2_info_freefrag_stats *stats, + unsigned int chunksize) +{ + if (chunksize > stats->ffs_max) + stats->ffs_max = chunksize; + + if (chunksize < stats->ffs_min) + stats->ffs_min = chunksize; + + stats->ffs_avg += chunksize; + stats->ffs_free_chunks_real++; +} + +static void ocfs2_info_update_ffg(struct ocfs2_info_freefrag *ffg, + unsigned int chunksize) +{ + o2ffg_update_histogram(&(ffg->iff_ffs.ffs_fc_hist), chunksize); + o2ffg_update_stats(&(ffg->iff_ffs), chunksize); +} + +static int ocfs2_info_freefrag_scan_chain(struct ocfs2_super *osb, + struct inode *gb_inode, + struct ocfs2_dinode *gb_dinode, + struct ocfs2_chain_rec *rec, + struct ocfs2_info_freefrag *ffg, + u32 chunks_in_group) +{ + int status = 0, used; + u64 blkno; + + struct buffer_head *bh = NULL; + struct ocfs2_group_desc *bg = NULL; + + unsigned int max_bits, num_clusters; + unsigned int offset = 0, cluster, chunk; + unsigned int chunk_free, last_chunksize = 0; + + if (!le32_to_cpu(rec->c_free)) + goto bail; + + do { + if (!bg) + blkno = le64_to_cpu(rec->c_blkno); + else + blkno = le64_to_cpu(bg->bg_next_group); + + if (bh) { + brelse(bh); + bh = NULL; + } + + if (o2info_coherent(&ffg->iff_req)) + status = ocfs2_read_group_descriptor(gb_inode, + gb_dinode, + blkno, &bh); + else + status = ocfs2_read_blocks_sync(osb, blkno, 1, &bh); + + if (status < 0) { + mlog(ML_ERROR, "Can't read the group descriptor # " + "%llu from device.", (unsigned long long)blkno); + status = -EIO; + goto bail; + } + + bg = (struct ocfs2_group_desc *)bh->b_data; + + if (!le16_to_cpu(bg->bg_free_bits_count)) + continue; + + max_bits = le16_to_cpu(bg->bg_bits); + offset = 0; + + for (chunk = 0; chunk < chunks_in_group; chunk++) { + /* + * last chunk may be not an entire one. + */ + if ((offset + ffg->iff_chunksize) > max_bits) + num_clusters = max_bits - offset; + else + num_clusters = ffg->iff_chunksize; + + chunk_free = 0; + for (cluster = 0; cluster < num_clusters; cluster++) { + used = ocfs2_test_bit(offset, + (unsigned long *)bg->bg_bitmap); + /* + * - chunk_free counts free clusters in #N chunk. + * - last_chunksize records the size(in) clusters + * for the last real free chunk being counted. + */ + if (!used) { + last_chunksize++; + chunk_free++; + } + + if (used && last_chunksize) { + ocfs2_info_update_ffg(ffg, + last_chunksize); + last_chunksize = 0; + } + + offset++; + } + + if (chunk_free == ffg->iff_chunksize) + ffg->iff_ffs.ffs_free_chunks++; + } + + /* + * need to update the info for last free chunk. + */ + if (last_chunksize) + ocfs2_info_update_ffg(ffg, last_chunksize); + + } while (le64_to_cpu(bg->bg_next_group)); + +bail: + brelse(bh); + + return status; +} + +static int ocfs2_info_freefrag_scan_bitmap(struct ocfs2_super *osb, + struct inode *gb_inode, u64 blkno, + struct ocfs2_info_freefrag *ffg) +{ + u32 chunks_in_group; + int status = 0, unlock = 0, i; + + struct buffer_head *bh = NULL; + struct ocfs2_chain_list *cl = NULL; + struct ocfs2_chain_rec *rec = NULL; + struct ocfs2_dinode *gb_dinode = NULL; + + if (gb_inode) + mutex_lock(&gb_inode->i_mutex); + + if (o2info_coherent(&ffg->iff_req)) { + status = ocfs2_inode_lock(gb_inode, &bh, 0); + if (status < 0) { + mlog_errno(status); + goto bail; + } + unlock = 1; + } else { + status = ocfs2_read_blocks_sync(osb, blkno, 1, &bh); + if (status < 0) { + mlog_errno(status); + goto bail; + } + } + + gb_dinode = (struct ocfs2_dinode *)bh->b_data; + cl = &(gb_dinode->id2.i_chain); + + /* + * Chunksize(in) clusters from userspace should be + * less than clusters in a group. + */ + if (ffg->iff_chunksize > le16_to_cpu(cl->cl_cpg)) { + status = -EINVAL; + goto bail; + } + + memset(&ffg->iff_ffs, 0, sizeof(struct ocfs2_info_freefrag_stats)); + + ffg->iff_ffs.ffs_min = ~0U; + ffg->iff_ffs.ffs_clusters = + le32_to_cpu(gb_dinode->id1.bitmap1.i_total); + ffg->iff_ffs.ffs_free_clusters = ffg->iff_ffs.ffs_clusters - + le32_to_cpu(gb_dinode->id1.bitmap1.i_used); + + chunks_in_group = le16_to_cpu(cl->cl_cpg) / ffg->iff_chunksize + 1; + + for (i = 0; i < le16_to_cpu(cl->cl_next_free_rec); i++) { + rec = &(cl->cl_recs[i]); + status = ocfs2_info_freefrag_scan_chain(osb, gb_inode, + gb_dinode, + rec, ffg, + chunks_in_group); + if (status) + goto bail; + } + + if (ffg->iff_ffs.ffs_free_chunks_real) + ffg->iff_ffs.ffs_avg = (ffg->iff_ffs.ffs_avg / + ffg->iff_ffs.ffs_free_chunks_real); +bail: + if (unlock) + ocfs2_inode_unlock(gb_inode, 0); + + if (gb_inode) + mutex_unlock(&gb_inode->i_mutex); + + if (gb_inode) + iput(gb_inode); + + brelse(bh); + + return status; +} + +static int ocfs2_info_handle_freefrag(struct inode *inode, + struct ocfs2_info_request __user *req) +{ + u64 blkno = -1; + char namebuf[40]; + int status = -EFAULT, type = GLOBAL_BITMAP_SYSTEM_INODE; + + struct ocfs2_info_freefrag *oiff; + struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); + struct inode *gb_inode = NULL; + + oiff = kzalloc(sizeof(struct ocfs2_info_freefrag), GFP_KERNEL); + if (!oiff) { + status = -ENOMEM; + mlog_errno(status); + goto out_err; + } + + if (o2info_from_user(*oiff, req)) + goto bail; + /* + * chunksize from userspace should be power of 2. + */ + if ((oiff->iff_chunksize & (oiff->iff_chunksize - 1)) || + (!oiff->iff_chunksize)) { + status = -EINVAL; + goto bail; + } + + if (o2info_coherent(&oiff->iff_req)) { + gb_inode = ocfs2_get_system_file_inode(osb, type, + OCFS2_INVALID_SLOT); + if (!gb_inode) { + mlog(ML_ERROR, "unable to get global_bitmap inode\n"); + status = -EIO; + goto bail; + } + } else { + ocfs2_sprintf_system_inode_name(namebuf, sizeof(namebuf), type, + OCFS2_INVALID_SLOT); + status = ocfs2_lookup_ino_from_name(osb->sys_root_inode, + namebuf, + strlen(namebuf), + &blkno); + if (status < 0) { + status = -ENOENT; + goto bail; + } + } + + status = ocfs2_info_freefrag_scan_bitmap(osb, gb_inode, blkno, oiff); + if (status < 0) + goto bail; + + o2info_set_request_filled(&oiff->iff_req); + + if (o2info_to_user(*oiff, req)) { + status = -EFAULT; + goto bail; + } + + status = 0; +bail: + if (status) + o2info_set_request_error(&oiff->iff_req, req); + kfree(oiff); +out_err: return status; } -int ocfs2_info_handle_unknown(struct inode *inode, - struct ocfs2_info_request __user *req) +static int ocfs2_info_handle_unknown(struct inode *inode, + struct ocfs2_info_request __user *req) { int status = -EFAULT; struct ocfs2_info_request oir; @@ -327,7 +733,7 @@ int ocfs2_info_handle_unknown(struct inode *inode, if (o2info_from_user(oir, req)) goto bail; - o2info_clear_request_filled(oir); + o2info_clear_request_filled(&oir); if (o2info_to_user(oir, req)) goto bail; @@ -335,7 +741,7 @@ int ocfs2_info_handle_unknown(struct inode *inode, status = 0; bail: if (status) - o2info_set_request_error(oir, req); + o2info_set_request_error(&oir, req); return status; } @@ -347,8 +753,8 @@ bail: * - distinguish different requests. * - validate size of different requests. */ -int ocfs2_info_handle_request(struct inode *inode, - struct ocfs2_info_request __user *req) +static int ocfs2_info_handle_request(struct inode *inode, + struct ocfs2_info_request __user *req) { int status = -EFAULT; struct ocfs2_info_request oir; @@ -389,6 +795,14 @@ int ocfs2_info_handle_request(struct inode *inode, if (oir.ir_size == sizeof(struct ocfs2_info_journal_size)) status = ocfs2_info_handle_journal_size(inode, req); break; + case OCFS2_INFO_FREEINODE: + if (oir.ir_size == sizeof(struct ocfs2_info_freeinode)) + status = ocfs2_info_handle_freeinode(inode, req); + break; + case OCFS2_INFO_FREEFRAG: + if (oir.ir_size == sizeof(struct ocfs2_info_freefrag)) + status = ocfs2_info_handle_freefrag(inode, req); + break; default: status = ocfs2_info_handle_unknown(inode, req); break; @@ -398,8 +812,8 @@ bail: return status; } -int ocfs2_get_request_ptr(struct ocfs2_info *info, int idx, - u64 *req_addr, int compat_flag) +static int ocfs2_get_request_ptr(struct ocfs2_info *info, int idx, + u64 *req_addr, int compat_flag) { int status = -EFAULT; u64 __user *bp = NULL; @@ -436,8 +850,8 @@ bail: * a better backward&forward compatibility, since a small piece of * request will be less likely to be broken if disk layout get changed. */ -int ocfs2_info_handle(struct inode *inode, struct ocfs2_info *info, - int compat_flag) +static int ocfs2_info_handle(struct inode *inode, struct ocfs2_info *info, + int compat_flag) { int i, status = 0; u64 req_addr; @@ -455,7 +869,7 @@ int ocfs2_info_handle(struct inode *inode, struct ocfs2_info *info, if (status) break; - reqp = (struct ocfs2_info_request *)(unsigned long)req_addr; + reqp = (struct ocfs2_info_request __user *)(unsigned long)req_addr; if (!reqp) { status = -EINVAL; goto bail; @@ -472,16 +886,18 @@ bail: long ocfs2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) { - struct inode *inode = filp->f_path.dentry->d_inode; + struct inode *inode = file_inode(filp); unsigned int flags; int new_clusters; int status; struct ocfs2_space_resv sr; struct ocfs2_new_group_input input; struct reflink_arguments args; - const char *old_path, *new_path; + const char __user *old_path; + const char __user *new_path; bool preserve; struct ocfs2_info info; + void __user *argp = (void __user *)arg; switch (cmd) { case OCFS2_IOC_GETFLAGS: @@ -495,12 +911,12 @@ long ocfs2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) if (get_user(flags, (int __user *) arg)) return -EFAULT; - status = mnt_want_write(filp->f_path.mnt); + status = mnt_want_write_file(filp); if (status) return status; status = ocfs2_set_inode_attr(inode, flags, OCFS2_FL_MODIFIABLE); - mnt_drop_write(filp->f_path.mnt); + mnt_drop_write_file(filp); return status; case OCFS2_IOC_RESVSP: case OCFS2_IOC_RESVSP64: @@ -517,7 +933,12 @@ long ocfs2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) if (get_user(new_clusters, (int __user *)arg)) return -EFAULT; - return ocfs2_group_extend(inode, new_clusters); + status = mnt_want_write_file(filp); + if (status) + return status; + status = ocfs2_group_extend(inode, new_clusters); + mnt_drop_write_file(filp); + return status; case OCFS2_IOC_GROUP_ADD: case OCFS2_IOC_GROUP_ADD64: if (!capable(CAP_SYS_RESOURCE)) @@ -526,22 +947,54 @@ long ocfs2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) if (copy_from_user(&input, (int __user *) arg, sizeof(input))) return -EFAULT; - return ocfs2_group_add(inode, &input); + status = mnt_want_write_file(filp); + if (status) + return status; + status = ocfs2_group_add(inode, &input); + mnt_drop_write_file(filp); + return status; case OCFS2_IOC_REFLINK: - if (copy_from_user(&args, (struct reflink_arguments *)arg, - sizeof(args))) + if (copy_from_user(&args, argp, sizeof(args))) return -EFAULT; - old_path = (const char *)(unsigned long)args.old_path; - new_path = (const char *)(unsigned long)args.new_path; + old_path = (const char __user *)(unsigned long)args.old_path; + new_path = (const char __user *)(unsigned long)args.new_path; preserve = (args.preserve != 0); return ocfs2_reflink_ioctl(inode, old_path, new_path, preserve); case OCFS2_IOC_INFO: - if (copy_from_user(&info, (struct ocfs2_info __user *)arg, - sizeof(struct ocfs2_info))) + if (copy_from_user(&info, argp, sizeof(struct ocfs2_info))) return -EFAULT; return ocfs2_info_handle(inode, &info, 0); + case FITRIM: + { + struct super_block *sb = inode->i_sb; + struct request_queue *q = bdev_get_queue(sb->s_bdev); + struct fstrim_range range; + int ret = 0; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + if (!blk_queue_discard(q)) + return -EOPNOTSUPP; + + if (copy_from_user(&range, argp, sizeof(range))) + return -EFAULT; + + range.minlen = max_t(u64, q->limits.discard_granularity, + range.minlen); + ret = ocfs2_trim_fs(sb, &range); + if (ret < 0) + return ret; + + if (copy_to_user(argp, &range, sizeof(range))) + return -EFAULT; + + return 0; + } + case OCFS2_IOC_MOVE_EXT: + return ocfs2_ioctl_move_extents(filp, argp); default: return -ENOTTY; } @@ -552,8 +1005,9 @@ long ocfs2_compat_ioctl(struct file *file, unsigned cmd, unsigned long arg) { bool preserve; struct reflink_arguments args; - struct inode *inode = file->f_path.dentry->d_inode; + struct inode *inode = file_inode(file); struct ocfs2_info info; + void __user *argp = (void __user *)arg; switch (cmd) { case OCFS2_IOC32_GETFLAGS: @@ -569,21 +1023,22 @@ long ocfs2_compat_ioctl(struct file *file, unsigned cmd, unsigned long arg) case OCFS2_IOC_GROUP_EXTEND: case OCFS2_IOC_GROUP_ADD: case OCFS2_IOC_GROUP_ADD64: + case FITRIM: break; case OCFS2_IOC_REFLINK: - if (copy_from_user(&args, (struct reflink_arguments *)arg, - sizeof(args))) + if (copy_from_user(&args, argp, sizeof(args))) return -EFAULT; preserve = (args.preserve != 0); return ocfs2_reflink_ioctl(inode, compat_ptr(args.old_path), compat_ptr(args.new_path), preserve); case OCFS2_IOC_INFO: - if (copy_from_user(&info, (struct ocfs2_info __user *)arg, - sizeof(struct ocfs2_info))) + if (copy_from_user(&info, argp, sizeof(struct ocfs2_info))) return -EFAULT; return ocfs2_info_handle(inode, &info, 1); + case OCFS2_IOC_MOVE_EXT: + break; default: return -ENOIOCTLCMD; } diff --git a/fs/ocfs2/journal.c b/fs/ocfs2/journal.c index b141a44605c..4b0c68849b3 100644 --- a/fs/ocfs2/journal.c +++ b/fs/ocfs2/journal.c @@ -30,6 +30,7 @@ #include <linux/kthread.h> #include <linux/time.h> #include <linux/random.h> +#include <linux/delay.h> #include <cluster/masklog.h> @@ -355,11 +356,14 @@ handle_t *ocfs2_start_trans(struct ocfs2_super *osb, int max_buffs) if (journal_current_handle()) return jbd2_journal_start(journal, max_buffs); + sb_start_intwrite(osb->sb); + down_read(&osb->journal->j_trans_barrier); handle = jbd2_journal_start(journal, max_buffs); if (IS_ERR(handle)) { up_read(&osb->journal->j_trans_barrier); + sb_end_intwrite(osb->sb); mlog_errno(PTR_ERR(handle)); @@ -388,8 +392,10 @@ int ocfs2_commit_trans(struct ocfs2_super *osb, if (ret < 0) mlog_errno(ret); - if (!nested) + if (!nested) { up_read(&journal->j_trans_barrier); + sb_end_intwrite(osb->sb); + } return ret; } @@ -450,6 +456,41 @@ bail: return status; } +/* + * If we have fewer than thresh credits, extend by OCFS2_MAX_TRANS_DATA. + * If that fails, restart the transaction & regain write access for the + * buffer head which is used for metadata modifications. + * Taken from Ext4: extend_or_restart_transaction() + */ +int ocfs2_allocate_extend_trans(handle_t *handle, int thresh) +{ + int status, old_nblks; + + BUG_ON(!handle); + + old_nblks = handle->h_buffer_credits; + trace_ocfs2_allocate_extend_trans(old_nblks, thresh); + + if (old_nblks < thresh) + return 0; + + status = jbd2_journal_extend(handle, OCFS2_MAX_TRANS_DATA); + if (status < 0) { + mlog_errno(status); + goto bail; + } + + if (status > 0) { + status = jbd2_journal_restart(handle, OCFS2_MAX_TRANS_DATA); + if (status < 0) + mlog_errno(status); + } + +bail: + return status; +} + + struct ocfs2_triggers { struct jbd2_buffer_trigger_type ot_triggers; int ot_offset; @@ -796,14 +837,14 @@ int ocfs2_journal_init(struct ocfs2_journal *journal, int *dirty) inode_lock = 1; di = (struct ocfs2_dinode *)bh->b_data; - if (inode->i_size < OCFS2_MIN_JOURNAL_SIZE) { + if (i_size_read(inode) < OCFS2_MIN_JOURNAL_SIZE) { mlog(ML_ERROR, "Journal file size (%lld) is too small!\n", - inode->i_size); + i_size_read(inode)); status = -EINVAL; goto done; } - trace_ocfs2_journal_init(inode->i_size, + trace_ocfs2_journal_init(i_size_read(inode), (unsigned long long)inode->i_blocks, OCFS2_I(inode)->ip_clusters); @@ -1091,7 +1132,7 @@ static int ocfs2_force_read_journal(struct inode *inode) memset(bhs, 0, sizeof(struct buffer_head *) * CONCURRENT_JOURNAL_FILL); - num_blocks = ocfs2_blocks_for_bytes(inode->i_sb, inode->i_size); + num_blocks = ocfs2_blocks_for_bytes(inode->i_sb, i_size_read(inode)); v_blkno = 0; while (v_blkno < num_blocks) { status = ocfs2_extent_map_get_blocks(inode, v_blkno, @@ -1229,11 +1270,8 @@ static void ocfs2_queue_recovery_completion(struct ocfs2_journal *journal, /* Though we wish to avoid it, we are in fact safe in * skipping local alloc cleanup as fsck.ocfs2 is more * than capable of reclaiming unused space. */ - if (la_dinode) - kfree(la_dinode); - - if (tl_dinode) - kfree(tl_dinode); + kfree(la_dinode); + kfree(tl_dinode); if (qrec) ocfs2_free_quota_recovery(qrec); @@ -1260,6 +1298,9 @@ void ocfs2_complete_mount_recovery(struct ocfs2_super *osb) { struct ocfs2_journal *journal = osb->journal; + if (ocfs2_is_hard_readonly(osb)) + return; + /* No need to queue up our truncate_log as regular cleanup will catch * that */ ocfs2_queue_recovery_completion(journal, osb->slot_num, @@ -1400,8 +1441,7 @@ bail: mutex_unlock(&osb->recovery_lock); - if (rm_quota) - kfree(rm_quota); + kfree(rm_quota); /* no one is callint kthread_stop() for us so the kthread() api * requires that we call do_exit(). And it isn't exported, but @@ -1541,9 +1581,9 @@ static int ocfs2_replay_journal(struct ocfs2_super *osb, /* we need to run complete recovery for offline orphan slots */ ocfs2_replay_map_set_state(osb, REPLAY_NEEDED); - mlog(ML_NOTICE, "Recovering node %d from slot %d on device (%u,%u)\n", - node_num, slot_num, - MAJOR(osb->sb->s_dev), MINOR(osb->sb->s_dev)); + printk(KERN_NOTICE "ocfs2: Begin replay journal (node %d, slot %d) on "\ + "device (%u,%u)\n", node_num, slot_num, MAJOR(osb->sb->s_dev), + MINOR(osb->sb->s_dev)); OCFS2_I(inode)->ip_clusters = le32_to_cpu(fe->i_clusters); @@ -1598,6 +1638,9 @@ static int ocfs2_replay_journal(struct ocfs2_super *osb, jbd2_journal_destroy(journal); + printk(KERN_NOTICE "ocfs2: End replay journal (node %d, slot %d) on "\ + "device (%u,%u)\n", node_num, slot_num, MAJOR(osb->sb->s_dev), + MINOR(osb->sb->s_dev)); done: /* drop the lock on this nodes journal */ if (got_lock) @@ -1805,6 +1848,20 @@ static inline unsigned long ocfs2_orphan_scan_timeout(void) * every slot, queuing a recovery of the slot on the ocfs2_wq thread. This * is done to catch any orphans that are left over in orphan directories. * + * It scans all slots, even ones that are in use. It does so to handle the + * case described below: + * + * Node 1 has an inode it was using. The dentry went away due to memory + * pressure. Node 1 closes the inode, but it's on the free list. The node + * has the open lock. + * Node 2 unlinks the inode. It grabs the dentry lock to notify others, + * but node 1 has no dentry and doesn't get the message. It trylocks the + * open lock, sees that another node has a PR, and does nothing. + * Later node 2 runs its orphan dir. It igets the inode, trylocks the + * open lock, sees the PR still, and does nothing. + * Basically, we have to trigger an orphan iput on node 1. The only way + * for this to happen is if node 1 runs node 2's orphan dir. + * * ocfs2_queue_orphan_scan gets called every ORPHAN_SCAN_SCHEDULE_TIMEOUT * seconds. It gets an EX lock on os_lockres and checks sequence number * stored in LVB. If the sequence number has changed, it means some other @@ -1920,6 +1977,7 @@ void ocfs2_orphan_scan_start(struct ocfs2_super *osb) } struct ocfs2_orphan_filldir_priv { + struct dir_context ctx; struct inode *head; struct ocfs2_super *osb; }; @@ -1956,11 +2014,11 @@ static int ocfs2_queue_orphans(struct ocfs2_super *osb, { int status; struct inode *orphan_dir_inode = NULL; - struct ocfs2_orphan_filldir_priv priv; - loff_t pos = 0; - - priv.osb = osb; - priv.head = *head; + struct ocfs2_orphan_filldir_priv priv = { + .ctx.actor = ocfs2_orphan_filldir, + .osb = osb, + .head = *head + }; orphan_dir_inode = ocfs2_get_system_file_inode(osb, ORPHAN_DIR_SYSTEM_INODE, @@ -1978,8 +2036,7 @@ static int ocfs2_queue_orphans(struct ocfs2_super *osb, goto out; } - status = ocfs2_dir_foreach(orphan_dir_inode, &pos, &priv, - ocfs2_orphan_filldir); + status = ocfs2_dir_foreach(orphan_dir_inode, &priv.ctx); if (status) { mlog_errno(status); goto out_cluster; @@ -2076,12 +2133,6 @@ static int ocfs2_recover_orphans(struct ocfs2_super *osb, iter = oi->ip_next_orphan; spin_lock(&oi->ip_lock); - /* The remote delete code may have set these on the - * assumption that the other node would wipe them - * successfully. If they are still in the node's - * orphan dir, we need to reset that state. */ - oi->ip_flags &= ~(OCFS2_INODE_DELETED|OCFS2_INODE_SKIP_DELETE); - /* Set the proper information to get us going into * ocfs2_delete_inode. */ oi->ip_flags |= OCFS2_INODE_MAYBE_ORPHANED; @@ -2135,8 +2186,20 @@ static int ocfs2_commit_thread(void *arg) || kthread_should_stop()); status = ocfs2_commit_cache(osb); - if (status < 0) - mlog_errno(status); + if (status < 0) { + static unsigned long abort_warn_time; + + /* Warn about this once per minute */ + if (printk_timed_ratelimit(&abort_warn_time, 60*HZ)) + mlog(ML_ERROR, "status = %d, journal is " + "already aborted.\n", status); + /* + * After ocfs2_commit_cache() fails, j_num_trans has a + * non-zero value. Sleep here to avoid a busy-wait + * loop. + */ + msleep_interruptible(1000); + } if (kthread_should_stop() && atomic_read(&journal->j_num_trans)){ mlog(ML_KTHREAD, diff --git a/fs/ocfs2/journal.h b/fs/ocfs2/journal.h index 68cf2f6d3c6..7f8cde94abf 100644 --- a/fs/ocfs2/journal.h +++ b/fs/ocfs2/journal.h @@ -200,7 +200,6 @@ void ocfs2_complete_quota_recovery(struct ocfs2_super *osb); static inline void ocfs2_start_checkpoint(struct ocfs2_super *osb) { - atomic_set(&osb->needs_checkpoint, 1); wake_up(&osb->checkpoint_event); } @@ -259,6 +258,17 @@ handle_t *ocfs2_start_trans(struct ocfs2_super *osb, int ocfs2_commit_trans(struct ocfs2_super *osb, handle_t *handle); int ocfs2_extend_trans(handle_t *handle, int nblocks); +int ocfs2_allocate_extend_trans(handle_t *handle, + int thresh); + +/* + * Define an arbitrary limit for the amount of data we will anticipate + * writing to any given transaction. For unbounded transactions such as + * fallocate(2) we can write more than this, but we always + * start off at the maximum transaction size and grow the transaction + * optimistically as we go. + */ +#define OCFS2_MAX_TRANS_DATA 64U /* * Create access is for when we get a newly created buffer and we're @@ -441,10 +451,11 @@ static inline int ocfs2_mknod_credits(struct super_block *sb, int is_dir, #define OCFS2_SIMPLE_DIR_EXTEND_CREDITS (2) /* file update (nlink, etc) + directory mtime/ctime + dir entry block + quota - * update on dir + index leaf + dx root update for free list */ + * update on dir + index leaf + dx root update for free list + + * previous dirblock update in the free list */ static inline int ocfs2_link_credits(struct super_block *sb) { - return 2*OCFS2_INODE_UPDATE_CREDITS + 3 + + return 2*OCFS2_INODE_UPDATE_CREDITS + 4 + ocfs2_quota_trans_credits(sb); } @@ -513,8 +524,7 @@ static inline int ocfs2_calc_dxi_expand_credits(struct super_block *sb) * the result may be wrong. */ static inline int ocfs2_calc_extend_credits(struct super_block *sb, - struct ocfs2_extent_list *root_el, - u32 bits_wanted) + struct ocfs2_extent_list *root_el) { int bitmap_blocks, sysfile_bitmap_blocks, extent_blocks; @@ -616,4 +626,15 @@ static inline int ocfs2_begin_ordered_truncate(struct inode *inode, new_size); } +static inline void ocfs2_update_inode_fsync_trans(handle_t *handle, + struct inode *inode, + int datasync) +{ + struct ocfs2_inode_info *oi = OCFS2_I(inode); + + oi->i_sync_tid = handle->h_transaction->t_tid; + if (datasync) + oi->i_datasync_tid = handle->h_transaction->t_tid; +} + #endif /* OCFS2_JOURNAL_H */ diff --git a/fs/ocfs2/localalloc.c b/fs/ocfs2/localalloc.c index 210c3523754..04401345562 100644 --- a/fs/ocfs2/localalloc.c +++ b/fs/ocfs2/localalloc.c @@ -476,8 +476,7 @@ out: if (local_alloc_inode) iput(local_alloc_inode); - if (alloc_copy) - kfree(alloc_copy); + kfree(alloc_copy); } /* @@ -534,7 +533,7 @@ int ocfs2_begin_local_alloc_recovery(struct ocfs2_super *osb, mlog_errno(status); bail: - if ((status < 0) && (*alloc_copy)) { + if (status < 0) { kfree(*alloc_copy); *alloc_copy = NULL; } @@ -782,16 +781,54 @@ bail: return status; } +int ocfs2_free_local_alloc_bits(struct ocfs2_super *osb, + handle_t *handle, + struct ocfs2_alloc_context *ac, + u32 bit_off, + u32 num_bits) +{ + int status, start; + u32 clear_bits; + struct inode *local_alloc_inode; + void *bitmap; + struct ocfs2_dinode *alloc; + struct ocfs2_local_alloc *la; + + BUG_ON(ac->ac_which != OCFS2_AC_USE_LOCAL); + + local_alloc_inode = ac->ac_inode; + alloc = (struct ocfs2_dinode *) osb->local_alloc_bh->b_data; + la = OCFS2_LOCAL_ALLOC(alloc); + + bitmap = la->la_bitmap; + start = bit_off - le32_to_cpu(la->la_bm_off); + clear_bits = num_bits; + + status = ocfs2_journal_access_di(handle, + INODE_CACHE(local_alloc_inode), + osb->local_alloc_bh, + OCFS2_JOURNAL_ACCESS_WRITE); + if (status < 0) { + mlog_errno(status); + goto bail; + } + + while (clear_bits--) + ocfs2_clear_bit(start++, bitmap); + + le32_add_cpu(&alloc->id1.bitmap1.i_used, -num_bits); + ocfs2_journal_dirty(handle, osb->local_alloc_bh); + +bail: + return status; +} + static u32 ocfs2_local_alloc_count_bits(struct ocfs2_dinode *alloc) { - int i; - u8 *buffer; - u32 count = 0; + u32 count; struct ocfs2_local_alloc *la = OCFS2_LOCAL_ALLOC(alloc); - buffer = la->la_bitmap; - for (i = 0; i < le16_to_cpu(la->la_size); i++) - count += hweight8(buffer[i]); + count = memweight(la->la_bitmap, le16_to_cpu(la->la_size)); trace_ocfs2_local_alloc_count_bits(count); return count; @@ -1087,7 +1124,7 @@ static int ocfs2_local_alloc_reserve_for_window(struct ocfs2_super *osb, } retry_enospc: - (*ac)->ac_bits_wanted = osb->local_alloc_default_bits; + (*ac)->ac_bits_wanted = osb->local_alloc_bits; status = ocfs2_reserve_cluster_bitmap_bits(osb, *ac); if (status == -ENOSPC) { if (ocfs2_recalc_la_window(osb, OCFS2_LA_EVENT_ENOSPC) == @@ -1159,7 +1196,7 @@ retry_enospc: OCFS2_LA_DISABLED) goto bail; - ac->ac_bits_wanted = osb->local_alloc_default_bits; + ac->ac_bits_wanted = osb->local_alloc_bits; status = ocfs2_claim_clusters(handle, ac, osb->local_alloc_bits, &cluster_off, @@ -1294,8 +1331,7 @@ bail: if (main_bm_inode) iput(main_bm_inode); - if (alloc_copy) - kfree(alloc_copy); + kfree(alloc_copy); if (ac) ocfs2_free_alloc_context(ac); diff --git a/fs/ocfs2/localalloc.h b/fs/ocfs2/localalloc.h index 1be9b586446..44a7d1fb2de 100644 --- a/fs/ocfs2/localalloc.h +++ b/fs/ocfs2/localalloc.h @@ -55,6 +55,12 @@ int ocfs2_claim_local_alloc_bits(struct ocfs2_super *osb, u32 *bit_off, u32 *num_bits); +int ocfs2_free_local_alloc_bits(struct ocfs2_super *osb, + handle_t *handle, + struct ocfs2_alloc_context *ac, + u32 bit_off, + u32 num_bits); + void ocfs2_local_alloc_seen_free_bits(struct ocfs2_super *osb, unsigned int num_clusters); void ocfs2_la_enable_worker(struct work_struct *work); diff --git a/fs/ocfs2/locks.c b/fs/ocfs2/locks.c index e57c804069e..6b6d092b099 100644 --- a/fs/ocfs2/locks.c +++ b/fs/ocfs2/locks.c @@ -82,6 +82,8 @@ static int ocfs2_do_flock(struct file *file, struct inode *inode, } ret = flock_lock_file_wait(file, fl); + if (ret) + ocfs2_file_unlock(file); out: mutex_unlock(&fp->fp_mutex); diff --git a/fs/ocfs2/mmap.c b/fs/ocfs2/mmap.c index 3e9393ca39e..10d66c75cec 100644 --- a/fs/ocfs2/mmap.c +++ b/fs/ocfs2/mmap.c @@ -61,8 +61,8 @@ static int ocfs2_fault(struct vm_area_struct *area, struct vm_fault *vmf) static int __ocfs2_page_mkwrite(struct file *file, struct buffer_head *di_bh, struct page *page) { - int ret; - struct inode *inode = file->f_path.dentry->d_inode; + int ret = VM_FAULT_NOPAGE; + struct inode *inode = file_inode(file); struct address_space *mapping = inode->i_mapping; loff_t pos = page_offset(page); unsigned int len = PAGE_CACHE_SIZE; @@ -71,32 +71,25 @@ static int __ocfs2_page_mkwrite(struct file *file, struct buffer_head *di_bh, void *fsdata; loff_t size = i_size_read(inode); - /* - * Another node might have truncated while we were waiting on - * cluster locks. - * We don't check size == 0 before the shift. This is borrowed - * from do_generic_file_read. - */ last_index = (size - 1) >> PAGE_CACHE_SHIFT; - if (unlikely(!size || page->index > last_index)) { - ret = -EINVAL; - goto out; - } /* - * The i_size check above doesn't catch the case where nodes - * truncated and then re-extended the file. We'll re-check the - * page mapping after taking the page lock inside of - * ocfs2_write_begin_nolock(). + * There are cases that lead to the page no longer bebongs to the + * mapping. + * 1) pagecache truncates locally due to memory pressure. + * 2) pagecache truncates when another is taking EX lock against + * inode lock. see ocfs2_data_convert_worker. + * + * The i_size check doesn't catch the case where nodes truncated and + * then re-extended the file. We'll re-check the page mapping after + * taking the page lock inside of ocfs2_write_begin_nolock(). + * + * Let VM retry with these cases. */ - if (!PageUptodate(page) || page->mapping != inode->i_mapping) { - /* - * the page has been umapped in ocfs2_data_downconvert_worker. - * So return 0 here and let VFS retry. - */ - ret = 0; + if ((page->mapping != inode->i_mapping) || + (!PageUptodate(page)) || + (page_offset(page) >= size)) goto out; - } /* * Call ocfs2_write_begin() and ocfs2_write_end() to take @@ -116,17 +109,21 @@ static int __ocfs2_page_mkwrite(struct file *file, struct buffer_head *di_bh, if (ret) { if (ret != -ENOSPC) mlog_errno(ret); + if (ret == -ENOMEM) + ret = VM_FAULT_OOM; + else + ret = VM_FAULT_SIGBUS; goto out; } - ret = ocfs2_write_end_nolock(mapping, pos, len, len, locked_page, - fsdata); - if (ret < 0) { - mlog_errno(ret); + if (!locked_page) { + ret = VM_FAULT_NOPAGE; goto out; } + ret = ocfs2_write_end_nolock(mapping, pos, len, len, locked_page, + fsdata); BUG_ON(ret != len); - ret = 0; + ret = VM_FAULT_LOCKED; out: return ret; } @@ -134,11 +131,12 @@ out: static int ocfs2_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) { struct page *page = vmf->page; - struct inode *inode = vma->vm_file->f_path.dentry->d_inode; + struct inode *inode = file_inode(vma->vm_file); struct buffer_head *di_bh = NULL; sigset_t oldset; int ret; + sb_start_pagefault(inode->i_sb); ocfs2_block_signals(&oldset); /* @@ -168,30 +166,29 @@ static int ocfs2_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) out: ocfs2_unblock_signals(&oldset); - if (ret) - ret = VM_FAULT_SIGBUS; + sb_end_pagefault(inode->i_sb); return ret; } static const struct vm_operations_struct ocfs2_file_vm_ops = { .fault = ocfs2_fault, .page_mkwrite = ocfs2_page_mkwrite, + .remap_pages = generic_file_remap_pages, }; int ocfs2_mmap(struct file *file, struct vm_area_struct *vma) { int ret = 0, lock_level = 0; - ret = ocfs2_inode_lock_atime(file->f_dentry->d_inode, - file->f_vfsmnt, &lock_level); + ret = ocfs2_inode_lock_atime(file_inode(file), + file->f_path.mnt, &lock_level); if (ret < 0) { mlog_errno(ret); goto out; } - ocfs2_inode_unlock(file->f_dentry->d_inode, lock_level); + ocfs2_inode_unlock(file_inode(file), lock_level); out: vma->vm_ops = &ocfs2_file_vm_ops; - vma->vm_flags |= VM_CAN_NONLINEAR; return 0; } diff --git a/fs/ocfs2/move_extents.c b/fs/ocfs2/move_extents.c new file mode 100644 index 00000000000..599eb4c4c8b --- /dev/null +++ b/fs/ocfs2/move_extents.c @@ -0,0 +1,1078 @@ +/* -*- mode: c; c-basic-offset: 8; -*- + * vim: noexpandtab sw=8 ts=8 sts=0: + * + * move_extents.c + * + * Copyright (C) 2011 Oracle. All rights reserved. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public + * License version 2 as published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + */ +#include <linux/fs.h> +#include <linux/types.h> +#include <linux/mount.h> +#include <linux/swap.h> + +#include <cluster/masklog.h> + +#include "ocfs2.h" +#include "ocfs2_ioctl.h" + +#include "alloc.h" +#include "aops.h" +#include "dlmglue.h" +#include "extent_map.h" +#include "inode.h" +#include "journal.h" +#include "suballoc.h" +#include "uptodate.h" +#include "super.h" +#include "dir.h" +#include "buffer_head_io.h" +#include "sysfile.h" +#include "refcounttree.h" +#include "move_extents.h" + +struct ocfs2_move_extents_context { + struct inode *inode; + struct file *file; + int auto_defrag; + int partial; + int credits; + u32 new_phys_cpos; + u32 clusters_moved; + u64 refcount_loc; + struct ocfs2_move_extents *range; + struct ocfs2_extent_tree et; + struct ocfs2_alloc_context *meta_ac; + struct ocfs2_alloc_context *data_ac; + struct ocfs2_cached_dealloc_ctxt dealloc; +}; + +static int __ocfs2_move_extent(handle_t *handle, + struct ocfs2_move_extents_context *context, + u32 cpos, u32 len, u32 p_cpos, u32 new_p_cpos, + int ext_flags) +{ + int ret = 0, index; + struct inode *inode = context->inode; + struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); + struct ocfs2_extent_rec *rec, replace_rec; + struct ocfs2_path *path = NULL; + struct ocfs2_extent_list *el; + u64 ino = ocfs2_metadata_cache_owner(context->et.et_ci); + u64 old_blkno = ocfs2_clusters_to_blocks(inode->i_sb, p_cpos); + + ret = ocfs2_duplicate_clusters_by_page(handle, inode, cpos, + p_cpos, new_p_cpos, len); + if (ret) { + mlog_errno(ret); + goto out; + } + + memset(&replace_rec, 0, sizeof(replace_rec)); + replace_rec.e_cpos = cpu_to_le32(cpos); + replace_rec.e_leaf_clusters = cpu_to_le16(len); + replace_rec.e_blkno = cpu_to_le64(ocfs2_clusters_to_blocks(inode->i_sb, + new_p_cpos)); + + path = ocfs2_new_path_from_et(&context->et); + if (!path) { + ret = -ENOMEM; + mlog_errno(ret); + goto out; + } + + ret = ocfs2_find_path(INODE_CACHE(inode), path, cpos); + if (ret) { + mlog_errno(ret); + goto out; + } + + el = path_leaf_el(path); + + index = ocfs2_search_extent_list(el, cpos); + if (index == -1 || index >= le16_to_cpu(el->l_next_free_rec)) { + ocfs2_error(inode->i_sb, + "Inode %llu has an extent at cpos %u which can no " + "longer be found.\n", + (unsigned long long)ino, cpos); + ret = -EROFS; + goto out; + } + + rec = &el->l_recs[index]; + + BUG_ON(ext_flags != rec->e_flags); + /* + * after moving/defraging to new location, the extent is not going + * to be refcounted anymore. + */ + replace_rec.e_flags = ext_flags & ~OCFS2_EXT_REFCOUNTED; + + ret = ocfs2_journal_access_di(handle, INODE_CACHE(inode), + context->et.et_root_bh, + OCFS2_JOURNAL_ACCESS_WRITE); + if (ret) { + mlog_errno(ret); + goto out; + } + + ret = ocfs2_split_extent(handle, &context->et, path, index, + &replace_rec, context->meta_ac, + &context->dealloc); + if (ret) { + mlog_errno(ret); + goto out; + } + + ocfs2_journal_dirty(handle, context->et.et_root_bh); + + context->new_phys_cpos = new_p_cpos; + + /* + * need I to append truncate log for old clusters? + */ + if (old_blkno) { + if (ext_flags & OCFS2_EXT_REFCOUNTED) + ret = ocfs2_decrease_refcount(inode, handle, + ocfs2_blocks_to_clusters(osb->sb, + old_blkno), + len, context->meta_ac, + &context->dealloc, 1); + else + ret = ocfs2_truncate_log_append(osb, handle, + old_blkno, len); + } + + ocfs2_update_inode_fsync_trans(handle, inode, 0); +out: + ocfs2_free_path(path); + return ret; +} + +/* + * lock allocators, and reserving appropriate number of bits for + * meta blocks and data clusters. + * + * in some cases, we don't need to reserve clusters, just let data_ac + * be NULL. + */ +static int ocfs2_lock_allocators_move_extents(struct inode *inode, + struct ocfs2_extent_tree *et, + u32 clusters_to_move, + u32 extents_to_split, + struct ocfs2_alloc_context **meta_ac, + struct ocfs2_alloc_context **data_ac, + int extra_blocks, + int *credits) +{ + int ret, num_free_extents; + unsigned int max_recs_needed = 2 * extents_to_split + clusters_to_move; + struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); + + num_free_extents = ocfs2_num_free_extents(osb, et); + if (num_free_extents < 0) { + ret = num_free_extents; + mlog_errno(ret); + goto out; + } + + if (!num_free_extents || + (ocfs2_sparse_alloc(osb) && num_free_extents < max_recs_needed)) + extra_blocks += ocfs2_extend_meta_needed(et->et_root_el); + + ret = ocfs2_reserve_new_metadata_blocks(osb, extra_blocks, meta_ac); + if (ret) { + mlog_errno(ret); + goto out; + } + + if (data_ac) { + ret = ocfs2_reserve_clusters(osb, clusters_to_move, data_ac); + if (ret) { + mlog_errno(ret); + goto out; + } + } + + *credits += ocfs2_calc_extend_credits(osb->sb, et->et_root_el); + + mlog(0, "reserve metadata_blocks: %d, data_clusters: %u, credits: %d\n", + extra_blocks, clusters_to_move, *credits); +out: + if (ret) { + if (*meta_ac) { + ocfs2_free_alloc_context(*meta_ac); + *meta_ac = NULL; + } + } + + return ret; +} + +/* + * Using one journal handle to guarantee the data consistency in case + * crash happens anywhere. + * + * XXX: defrag can end up with finishing partial extent as requested, + * due to not enough contiguous clusters can be found in allocator. + */ +static int ocfs2_defrag_extent(struct ocfs2_move_extents_context *context, + u32 cpos, u32 phys_cpos, u32 *len, int ext_flags) +{ + int ret, credits = 0, extra_blocks = 0, partial = context->partial; + handle_t *handle; + struct inode *inode = context->inode; + struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); + struct inode *tl_inode = osb->osb_tl_inode; + struct ocfs2_refcount_tree *ref_tree = NULL; + u32 new_phys_cpos, new_len; + u64 phys_blkno = ocfs2_clusters_to_blocks(inode->i_sb, phys_cpos); + + if ((ext_flags & OCFS2_EXT_REFCOUNTED) && *len) { + + BUG_ON(!(OCFS2_I(inode)->ip_dyn_features & + OCFS2_HAS_REFCOUNT_FL)); + + BUG_ON(!context->refcount_loc); + + ret = ocfs2_lock_refcount_tree(osb, context->refcount_loc, 1, + &ref_tree, NULL); + if (ret) { + mlog_errno(ret); + return ret; + } + + ret = ocfs2_prepare_refcount_change_for_del(inode, + context->refcount_loc, + phys_blkno, + *len, + &credits, + &extra_blocks); + if (ret) { + mlog_errno(ret); + goto out; + } + } + + ret = ocfs2_lock_allocators_move_extents(inode, &context->et, *len, 1, + &context->meta_ac, + &context->data_ac, + extra_blocks, &credits); + if (ret) { + mlog_errno(ret); + goto out; + } + + /* + * should be using allocation reservation strategy there? + * + * if (context->data_ac) + * context->data_ac->ac_resv = &OCFS2_I(inode)->ip_la_data_resv; + */ + + mutex_lock(&tl_inode->i_mutex); + + if (ocfs2_truncate_log_needs_flush(osb)) { + ret = __ocfs2_flush_truncate_log(osb); + if (ret < 0) { + mlog_errno(ret); + goto out_unlock_mutex; + } + } + + handle = ocfs2_start_trans(osb, credits); + if (IS_ERR(handle)) { + ret = PTR_ERR(handle); + mlog_errno(ret); + goto out_unlock_mutex; + } + + ret = __ocfs2_claim_clusters(handle, context->data_ac, 1, *len, + &new_phys_cpos, &new_len); + if (ret) { + mlog_errno(ret); + goto out_commit; + } + + /* + * allowing partial extent moving is kind of 'pros and cons', it makes + * whole defragmentation less likely to fail, on the contrary, the bad + * thing is it may make the fs even more fragmented after moving, let + * userspace make a good decision here. + */ + if (new_len != *len) { + mlog(0, "len_claimed: %u, len: %u\n", new_len, *len); + if (!partial) { + context->range->me_flags &= ~OCFS2_MOVE_EXT_FL_COMPLETE; + ret = -ENOSPC; + goto out_commit; + } + } + + mlog(0, "cpos: %u, phys_cpos: %u, new_phys_cpos: %u\n", cpos, + phys_cpos, new_phys_cpos); + + ret = __ocfs2_move_extent(handle, context, cpos, new_len, phys_cpos, + new_phys_cpos, ext_flags); + if (ret) + mlog_errno(ret); + + if (partial && (new_len != *len)) + *len = new_len; + + /* + * Here we should write the new page out first if we are + * in write-back mode. + */ + ret = ocfs2_cow_sync_writeback(inode->i_sb, context->inode, cpos, *len); + if (ret) + mlog_errno(ret); + +out_commit: + ocfs2_commit_trans(osb, handle); + +out_unlock_mutex: + mutex_unlock(&tl_inode->i_mutex); + + if (context->data_ac) { + ocfs2_free_alloc_context(context->data_ac); + context->data_ac = NULL; + } + + if (context->meta_ac) { + ocfs2_free_alloc_context(context->meta_ac); + context->meta_ac = NULL; + } + +out: + if (ref_tree) + ocfs2_unlock_refcount_tree(osb, ref_tree, 1); + + return ret; +} + +/* + * find the victim alloc group, where #blkno fits. + */ +static int ocfs2_find_victim_alloc_group(struct inode *inode, + u64 vict_blkno, + int type, int slot, + int *vict_bit, + struct buffer_head **ret_bh) +{ + int ret, i, bits_per_unit = 0; + u64 blkno; + char namebuf[40]; + + struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); + struct buffer_head *ac_bh = NULL, *gd_bh = NULL; + struct ocfs2_chain_list *cl; + struct ocfs2_chain_rec *rec; + struct ocfs2_dinode *ac_dinode; + struct ocfs2_group_desc *bg; + + ocfs2_sprintf_system_inode_name(namebuf, sizeof(namebuf), type, slot); + ret = ocfs2_lookup_ino_from_name(osb->sys_root_inode, namebuf, + strlen(namebuf), &blkno); + if (ret) { + ret = -ENOENT; + goto out; + } + + ret = ocfs2_read_blocks_sync(osb, blkno, 1, &ac_bh); + if (ret) { + mlog_errno(ret); + goto out; + } + + ac_dinode = (struct ocfs2_dinode *)ac_bh->b_data; + cl = &(ac_dinode->id2.i_chain); + rec = &(cl->cl_recs[0]); + + if (type == GLOBAL_BITMAP_SYSTEM_INODE) + bits_per_unit = osb->s_clustersize_bits - + inode->i_sb->s_blocksize_bits; + /* + * 'vict_blkno' was out of the valid range. + */ + if ((vict_blkno < le64_to_cpu(rec->c_blkno)) || + (vict_blkno >= (le32_to_cpu(ac_dinode->id1.bitmap1.i_total) << + bits_per_unit))) { + ret = -EINVAL; + goto out; + } + + for (i = 0; i < le16_to_cpu(cl->cl_next_free_rec); i++) { + + rec = &(cl->cl_recs[i]); + if (!rec) + continue; + + bg = NULL; + + do { + if (!bg) + blkno = le64_to_cpu(rec->c_blkno); + else + blkno = le64_to_cpu(bg->bg_next_group); + + if (gd_bh) { + brelse(gd_bh); + gd_bh = NULL; + } + + ret = ocfs2_read_blocks_sync(osb, blkno, 1, &gd_bh); + if (ret) { + mlog_errno(ret); + goto out; + } + + bg = (struct ocfs2_group_desc *)gd_bh->b_data; + + if (vict_blkno < (le64_to_cpu(bg->bg_blkno) + + le16_to_cpu(bg->bg_bits))) { + + *ret_bh = gd_bh; + *vict_bit = (vict_blkno - blkno) >> + bits_per_unit; + mlog(0, "find the victim group: #%llu, " + "total_bits: %u, vict_bit: %u\n", + blkno, le16_to_cpu(bg->bg_bits), + *vict_bit); + goto out; + } + + } while (le64_to_cpu(bg->bg_next_group)); + } + + ret = -EINVAL; +out: + brelse(ac_bh); + + /* + * caller has to release the gd_bh properly. + */ + return ret; +} + +/* + * XXX: helper to validate and adjust moving goal. + */ +static int ocfs2_validate_and_adjust_move_goal(struct inode *inode, + struct ocfs2_move_extents *range) +{ + int ret, goal_bit = 0; + + struct buffer_head *gd_bh = NULL; + struct ocfs2_group_desc *bg; + struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); + int c_to_b = 1 << (osb->s_clustersize_bits - + inode->i_sb->s_blocksize_bits); + + /* + * make goal become cluster aligned. + */ + range->me_goal = ocfs2_block_to_cluster_start(inode->i_sb, + range->me_goal); + /* + * validate goal sits within global_bitmap, and return the victim + * group desc + */ + ret = ocfs2_find_victim_alloc_group(inode, range->me_goal, + GLOBAL_BITMAP_SYSTEM_INODE, + OCFS2_INVALID_SLOT, + &goal_bit, &gd_bh); + if (ret) + goto out; + + bg = (struct ocfs2_group_desc *)gd_bh->b_data; + + /* + * moving goal is not allowd to start with a group desc blok(#0 blk) + * let's compromise to the latter cluster. + */ + if (range->me_goal == le64_to_cpu(bg->bg_blkno)) + range->me_goal += c_to_b; + + /* + * movement is not gonna cross two groups. + */ + if ((le16_to_cpu(bg->bg_bits) - goal_bit) * osb->s_clustersize < + range->me_len) { + ret = -EINVAL; + goto out; + } + /* + * more exact validations/adjustments will be performed later during + * moving operation for each extent range. + */ + mlog(0, "extents get ready to be moved to #%llu block\n", + range->me_goal); + +out: + brelse(gd_bh); + + return ret; +} + +static void ocfs2_probe_alloc_group(struct inode *inode, struct buffer_head *bh, + int *goal_bit, u32 move_len, u32 max_hop, + u32 *phys_cpos) +{ + int i, used, last_free_bits = 0, base_bit = *goal_bit; + struct ocfs2_group_desc *gd = (struct ocfs2_group_desc *)bh->b_data; + u32 base_cpos = ocfs2_blocks_to_clusters(inode->i_sb, + le64_to_cpu(gd->bg_blkno)); + + for (i = base_bit; i < le16_to_cpu(gd->bg_bits); i++) { + + used = ocfs2_test_bit(i, (unsigned long *)gd->bg_bitmap); + if (used) { + /* + * we even tried searching the free chunk by jumping + * a 'max_hop' distance, but still failed. + */ + if ((i - base_bit) > max_hop) { + *phys_cpos = 0; + break; + } + + if (last_free_bits) + last_free_bits = 0; + + continue; + } else + last_free_bits++; + + if (last_free_bits == move_len) { + *goal_bit = i; + *phys_cpos = base_cpos + i; + break; + } + } + + mlog(0, "found phys_cpos: %u to fit the wanted moving.\n", *phys_cpos); +} + +static int ocfs2_move_extent(struct ocfs2_move_extents_context *context, + u32 cpos, u32 phys_cpos, u32 *new_phys_cpos, + u32 len, int ext_flags) +{ + int ret, credits = 0, extra_blocks = 0, goal_bit = 0; + handle_t *handle; + struct inode *inode = context->inode; + struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); + struct inode *tl_inode = osb->osb_tl_inode; + struct inode *gb_inode = NULL; + struct buffer_head *gb_bh = NULL; + struct buffer_head *gd_bh = NULL; + struct ocfs2_group_desc *gd; + struct ocfs2_refcount_tree *ref_tree = NULL; + u32 move_max_hop = ocfs2_blocks_to_clusters(inode->i_sb, + context->range->me_threshold); + u64 phys_blkno, new_phys_blkno; + + phys_blkno = ocfs2_clusters_to_blocks(inode->i_sb, phys_cpos); + + if ((ext_flags & OCFS2_EXT_REFCOUNTED) && len) { + + BUG_ON(!(OCFS2_I(inode)->ip_dyn_features & + OCFS2_HAS_REFCOUNT_FL)); + + BUG_ON(!context->refcount_loc); + + ret = ocfs2_lock_refcount_tree(osb, context->refcount_loc, 1, + &ref_tree, NULL); + if (ret) { + mlog_errno(ret); + return ret; + } + + ret = ocfs2_prepare_refcount_change_for_del(inode, + context->refcount_loc, + phys_blkno, + len, + &credits, + &extra_blocks); + if (ret) { + mlog_errno(ret); + goto out; + } + } + + ret = ocfs2_lock_allocators_move_extents(inode, &context->et, len, 1, + &context->meta_ac, + NULL, extra_blocks, &credits); + if (ret) { + mlog_errno(ret); + goto out; + } + + /* + * need to count 2 extra credits for global_bitmap inode and + * group descriptor. + */ + credits += OCFS2_INODE_UPDATE_CREDITS + 1; + + /* + * ocfs2_move_extent() didn't reserve any clusters in lock_allocators() + * logic, while we still need to lock the global_bitmap. + */ + gb_inode = ocfs2_get_system_file_inode(osb, GLOBAL_BITMAP_SYSTEM_INODE, + OCFS2_INVALID_SLOT); + if (!gb_inode) { + mlog(ML_ERROR, "unable to get global_bitmap inode\n"); + ret = -EIO; + goto out; + } + + mutex_lock(&gb_inode->i_mutex); + + ret = ocfs2_inode_lock(gb_inode, &gb_bh, 1); + if (ret) { + mlog_errno(ret); + goto out_unlock_gb_mutex; + } + + mutex_lock(&tl_inode->i_mutex); + + handle = ocfs2_start_trans(osb, credits); + if (IS_ERR(handle)) { + ret = PTR_ERR(handle); + mlog_errno(ret); + goto out_unlock_tl_inode; + } + + new_phys_blkno = ocfs2_clusters_to_blocks(inode->i_sb, *new_phys_cpos); + ret = ocfs2_find_victim_alloc_group(inode, new_phys_blkno, + GLOBAL_BITMAP_SYSTEM_INODE, + OCFS2_INVALID_SLOT, + &goal_bit, &gd_bh); + if (ret) { + mlog_errno(ret); + goto out_commit; + } + + /* + * probe the victim cluster group to find a proper + * region to fit wanted movement, it even will perfrom + * a best-effort attempt by compromising to a threshold + * around the goal. + */ + ocfs2_probe_alloc_group(inode, gd_bh, &goal_bit, len, move_max_hop, + new_phys_cpos); + if (!*new_phys_cpos) { + ret = -ENOSPC; + goto out_commit; + } + + ret = __ocfs2_move_extent(handle, context, cpos, len, phys_cpos, + *new_phys_cpos, ext_flags); + if (ret) { + mlog_errno(ret); + goto out_commit; + } + + gd = (struct ocfs2_group_desc *)gd_bh->b_data; + ret = ocfs2_alloc_dinode_update_counts(gb_inode, handle, gb_bh, len, + le16_to_cpu(gd->bg_chain)); + if (ret) { + mlog_errno(ret); + goto out_commit; + } + + ret = ocfs2_block_group_set_bits(handle, gb_inode, gd, gd_bh, + goal_bit, len); + if (ret) { + ocfs2_rollback_alloc_dinode_counts(gb_inode, gb_bh, len, + le16_to_cpu(gd->bg_chain)); + mlog_errno(ret); + } + + /* + * Here we should write the new page out first if we are + * in write-back mode. + */ + ret = ocfs2_cow_sync_writeback(inode->i_sb, context->inode, cpos, len); + if (ret) + mlog_errno(ret); + +out_commit: + ocfs2_commit_trans(osb, handle); + brelse(gd_bh); + +out_unlock_tl_inode: + mutex_unlock(&tl_inode->i_mutex); + + ocfs2_inode_unlock(gb_inode, 1); +out_unlock_gb_mutex: + mutex_unlock(&gb_inode->i_mutex); + brelse(gb_bh); + iput(gb_inode); + +out: + if (context->meta_ac) { + ocfs2_free_alloc_context(context->meta_ac); + context->meta_ac = NULL; + } + + if (ref_tree) + ocfs2_unlock_refcount_tree(osb, ref_tree, 1); + + return ret; +} + +/* + * Helper to calculate the defraging length in one run according to threshold. + */ +static void ocfs2_calc_extent_defrag_len(u32 *alloc_size, u32 *len_defraged, + u32 threshold, int *skip) +{ + if ((*alloc_size + *len_defraged) < threshold) { + /* + * proceed defragmentation until we meet the thresh + */ + *len_defraged += *alloc_size; + } else if (*len_defraged == 0) { + /* + * XXX: skip a large extent. + */ + *skip = 1; + } else { + /* + * split this extent to coalesce with former pieces as + * to reach the threshold. + * + * we're done here with one cycle of defragmentation + * in a size of 'thresh', resetting 'len_defraged' + * forces a new defragmentation. + */ + *alloc_size = threshold - *len_defraged; + *len_defraged = 0; + } +} + +static int __ocfs2_move_extents_range(struct buffer_head *di_bh, + struct ocfs2_move_extents_context *context) +{ + int ret = 0, flags, do_defrag, skip = 0; + u32 cpos, phys_cpos, move_start, len_to_move, alloc_size; + u32 len_defraged = 0, defrag_thresh = 0, new_phys_cpos = 0; + + struct inode *inode = context->inode; + struct ocfs2_dinode *di = (struct ocfs2_dinode *)di_bh->b_data; + struct ocfs2_move_extents *range = context->range; + struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); + + if ((i_size_read(inode) == 0) || (range->me_len == 0)) + return 0; + + if (OCFS2_I(inode)->ip_dyn_features & OCFS2_INLINE_DATA_FL) + return 0; + + context->refcount_loc = le64_to_cpu(di->i_refcount_loc); + + ocfs2_init_dinode_extent_tree(&context->et, INODE_CACHE(inode), di_bh); + ocfs2_init_dealloc_ctxt(&context->dealloc); + + /* + * TO-DO XXX: + * + * - xattr extents. + */ + + do_defrag = context->auto_defrag; + + /* + * extents moving happens in unit of clusters, for the sake + * of simplicity, we may ignore two clusters where 'byte_start' + * and 'byte_start + len' were within. + */ + move_start = ocfs2_clusters_for_bytes(osb->sb, range->me_start); + len_to_move = (range->me_start + range->me_len) >> + osb->s_clustersize_bits; + if (len_to_move >= move_start) + len_to_move -= move_start; + else + len_to_move = 0; + + if (do_defrag) { + defrag_thresh = range->me_threshold >> osb->s_clustersize_bits; + if (defrag_thresh <= 1) + goto done; + } else + new_phys_cpos = ocfs2_blocks_to_clusters(inode->i_sb, + range->me_goal); + + mlog(0, "Inode: %llu, start: %llu, len: %llu, cstart: %u, clen: %u, " + "thresh: %u\n", + (unsigned long long)OCFS2_I(inode)->ip_blkno, + (unsigned long long)range->me_start, + (unsigned long long)range->me_len, + move_start, len_to_move, defrag_thresh); + + cpos = move_start; + while (len_to_move) { + ret = ocfs2_get_clusters(inode, cpos, &phys_cpos, &alloc_size, + &flags); + if (ret) { + mlog_errno(ret); + goto out; + } + + if (alloc_size > len_to_move) + alloc_size = len_to_move; + + /* + * XXX: how to deal with a hole: + * + * - skip the hole of course + * - force a new defragmentation + */ + if (!phys_cpos) { + if (do_defrag) + len_defraged = 0; + + goto next; + } + + if (do_defrag) { + ocfs2_calc_extent_defrag_len(&alloc_size, &len_defraged, + defrag_thresh, &skip); + /* + * skip large extents + */ + if (skip) { + skip = 0; + goto next; + } + + mlog(0, "#Defrag: cpos: %u, phys_cpos: %u, " + "alloc_size: %u, len_defraged: %u\n", + cpos, phys_cpos, alloc_size, len_defraged); + + ret = ocfs2_defrag_extent(context, cpos, phys_cpos, + &alloc_size, flags); + } else { + ret = ocfs2_move_extent(context, cpos, phys_cpos, + &new_phys_cpos, alloc_size, + flags); + + new_phys_cpos += alloc_size; + } + + if (ret < 0) { + mlog_errno(ret); + goto out; + } + + context->clusters_moved += alloc_size; +next: + cpos += alloc_size; + len_to_move -= alloc_size; + } + +done: + range->me_flags |= OCFS2_MOVE_EXT_FL_COMPLETE; + +out: + range->me_moved_len = ocfs2_clusters_to_bytes(osb->sb, + context->clusters_moved); + range->me_new_offset = ocfs2_clusters_to_bytes(osb->sb, + context->new_phys_cpos); + + ocfs2_schedule_truncate_log_flush(osb, 1); + ocfs2_run_deallocs(osb, &context->dealloc); + + return ret; +} + +static int ocfs2_move_extents(struct ocfs2_move_extents_context *context) +{ + int status; + handle_t *handle; + struct inode *inode = context->inode; + struct ocfs2_dinode *di; + struct buffer_head *di_bh = NULL; + struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); + + if (!inode) + return -ENOENT; + + if (ocfs2_is_hard_readonly(osb) || ocfs2_is_soft_readonly(osb)) + return -EROFS; + + mutex_lock(&inode->i_mutex); + + /* + * This prevents concurrent writes from other nodes + */ + status = ocfs2_rw_lock(inode, 1); + if (status) { + mlog_errno(status); + goto out; + } + + status = ocfs2_inode_lock(inode, &di_bh, 1); + if (status) { + mlog_errno(status); + goto out_rw_unlock; + } + + /* + * rememer ip_xattr_sem also needs to be held if necessary + */ + down_write(&OCFS2_I(inode)->ip_alloc_sem); + + status = __ocfs2_move_extents_range(di_bh, context); + + up_write(&OCFS2_I(inode)->ip_alloc_sem); + if (status) { + mlog_errno(status); + goto out_inode_unlock; + } + + /* + * We update ctime for these changes + */ + handle = ocfs2_start_trans(osb, OCFS2_INODE_UPDATE_CREDITS); + if (IS_ERR(handle)) { + status = PTR_ERR(handle); + mlog_errno(status); + goto out_inode_unlock; + } + + status = ocfs2_journal_access_di(handle, INODE_CACHE(inode), di_bh, + OCFS2_JOURNAL_ACCESS_WRITE); + if (status) { + mlog_errno(status); + goto out_commit; + } + + di = (struct ocfs2_dinode *)di_bh->b_data; + inode->i_ctime = CURRENT_TIME; + di->i_ctime = cpu_to_le64(inode->i_ctime.tv_sec); + di->i_ctime_nsec = cpu_to_le32(inode->i_ctime.tv_nsec); + ocfs2_update_inode_fsync_trans(handle, inode, 0); + + ocfs2_journal_dirty(handle, di_bh); + +out_commit: + ocfs2_commit_trans(osb, handle); + +out_inode_unlock: + brelse(di_bh); + ocfs2_inode_unlock(inode, 1); +out_rw_unlock: + ocfs2_rw_unlock(inode, 1); +out: + mutex_unlock(&inode->i_mutex); + + return status; +} + +int ocfs2_ioctl_move_extents(struct file *filp, void __user *argp) +{ + int status; + + struct inode *inode = file_inode(filp); + struct ocfs2_move_extents range; + struct ocfs2_move_extents_context *context; + + if (!argp) + return -EINVAL; + + status = mnt_want_write_file(filp); + if (status) + return status; + + if ((!S_ISREG(inode->i_mode)) || !(filp->f_mode & FMODE_WRITE)) { + status = -EPERM; + goto out_drop; + } + + if (inode->i_flags & (S_IMMUTABLE|S_APPEND)) { + status = -EPERM; + goto out_drop; + } + + context = kzalloc(sizeof(struct ocfs2_move_extents_context), GFP_NOFS); + if (!context) { + status = -ENOMEM; + mlog_errno(status); + goto out_drop; + } + + context->inode = inode; + context->file = filp; + + if (copy_from_user(&range, argp, sizeof(range))) { + status = -EFAULT; + goto out_free; + } + + if (range.me_start > i_size_read(inode)) { + status = -EINVAL; + goto out_free; + } + + if (range.me_start + range.me_len > i_size_read(inode)) + range.me_len = i_size_read(inode) - range.me_start; + + context->range = ⦥ + + if (range.me_flags & OCFS2_MOVE_EXT_FL_AUTO_DEFRAG) { + context->auto_defrag = 1; + /* + * ok, the default theshold for the defragmentation + * is 1M, since our maximum clustersize was 1M also. + * any thought? + */ + if (!range.me_threshold) + range.me_threshold = 1024 * 1024; + + if (range.me_threshold > i_size_read(inode)) + range.me_threshold = i_size_read(inode); + + if (range.me_flags & OCFS2_MOVE_EXT_FL_PART_DEFRAG) + context->partial = 1; + } else { + /* + * first best-effort attempt to validate and adjust the goal + * (physical address in block), while it can't guarantee later + * operation can succeed all the time since global_bitmap may + * change a bit over time. + */ + + status = ocfs2_validate_and_adjust_move_goal(inode, &range); + if (status) + goto out_copy; + } + + status = ocfs2_move_extents(context); + if (status) + mlog_errno(status); +out_copy: + /* + * movement/defragmentation may end up being partially completed, + * that's the reason why we need to return userspace the finished + * length and new_offset even if failure happens somewhere. + */ + if (copy_to_user(argp, &range, sizeof(range))) + status = -EFAULT; + +out_free: + kfree(context); +out_drop: + mnt_drop_write_file(filp); + + return status; +} diff --git a/fs/ocfs2/move_extents.h b/fs/ocfs2/move_extents.h new file mode 100644 index 00000000000..4e143e81144 --- /dev/null +++ b/fs/ocfs2/move_extents.h @@ -0,0 +1,22 @@ +/* -*- mode: c; c-basic-offset: 8; -*- + * vim: noexpandtab sw=8 ts=8 sts=0: + * + * move_extents.h + * + * Copyright (C) 2011 Oracle. All rights reserved. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public + * License version 2 as published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + */ +#ifndef OCFS2_MOVE_EXTENTS_H +#define OCFS2_MOVE_EXTENTS_H + +int ocfs2_ioctl_move_extents(struct file *filp, void __user *argp); + +#endif /* OCFS2_MOVE_EXTENTS_H */ diff --git a/fs/ocfs2/namei.c b/fs/ocfs2/namei.c index e5d738cd9cc..8add6f1030d 100644 --- a/fs/ocfs2/namei.c +++ b/fs/ocfs2/namei.c @@ -98,7 +98,7 @@ static int ocfs2_create_symlink_data(struct ocfs2_super *osb, #define OCFS2_ORPHAN_NAMELEN ((int)(2 * sizeof(u64))) static struct dentry *ocfs2_lookup(struct inode *dir, struct dentry *dentry, - struct nameidata *nd) + unsigned int flags) { int status; u64 blkno; @@ -185,7 +185,7 @@ bail: return ret; } -static struct inode *ocfs2_get_init_inode(struct inode *dir, int mode) +static struct inode *ocfs2_get_init_inode(struct inode *dir, umode_t mode) { struct inode *inode; @@ -199,17 +199,30 @@ static struct inode *ocfs2_get_init_inode(struct inode *dir, int mode) * these are used by the support functions here and in * callers. */ if (S_ISDIR(mode)) - inode->i_nlink = 2; - else - inode->i_nlink = 1; + set_nlink(inode, 2); inode_init_owner(inode, dir, mode); dquot_initialize(inode); return inode; } +static void ocfs2_cleanup_add_entry_failure(struct ocfs2_super *osb, + struct dentry *dentry, struct inode *inode) +{ + struct ocfs2_dentry_lock *dl = dentry->d_fsdata; + + ocfs2_simple_drop_lockres(osb, &dl->dl_lockres); + ocfs2_lock_res_free(&dl->dl_lockres); + BUG_ON(dl->dl_count != 1); + spin_lock(&dentry_attach_lock); + dentry->d_fsdata = NULL; + spin_unlock(&dentry_attach_lock); + kfree(dl); + iput(inode); +} + static int ocfs2_mknod(struct inode *dir, struct dentry *dentry, - int mode, + umode_t mode, dev_t dev) { int status = 0; @@ -232,6 +245,8 @@ static int ocfs2_mknod(struct inode *dir, struct ocfs2_dir_lookup_result lookup = { NULL, }; sigset_t oldset; int did_block_signals = 0; + struct posix_acl *default_acl = NULL, *acl = NULL; + struct ocfs2_dentry_lock *dl = NULL; trace_ocfs2_mknod(dir, dentry, dentry->d_name.len, dentry->d_name.name, (unsigned long long)OCFS2_I(dir)->ip_blkno, @@ -333,6 +348,12 @@ static int ocfs2_mknod(struct inode *dir, goto leave; } + status = posix_acl_create(dir, &mode, &default_acl, &acl); + if (status) { + mlog_errno(status); + goto leave; + } + handle = ocfs2_start_trans(osb, ocfs2_mknod_credits(osb->sb, S_ISDIR(mode), xattr_credits)); @@ -381,8 +402,17 @@ static int ocfs2_mknod(struct inode *dir, inc_nlink(dir); } - status = ocfs2_init_acl(handle, inode, dir, new_fe_bh, parent_fe_bh, - meta_ac, data_ac); + if (default_acl) { + status = ocfs2_set_acl(handle, inode, new_fe_bh, + ACL_TYPE_DEFAULT, default_acl, + meta_ac, data_ac); + } + if (!status && acl) { + status = ocfs2_set_acl(handle, inode, new_fe_bh, + ACL_TYPE_ACCESS, acl, + meta_ac, data_ac); + } + if (status < 0) { mlog_errno(status); goto leave; @@ -409,6 +439,8 @@ static int ocfs2_mknod(struct inode *dir, goto leave; } + dl = dentry->d_fsdata; + status = ocfs2_add_entry(handle, dentry, inode, OCFS2_I(inode)->ip_blkno, parent_fe_bh, &lookup); @@ -421,6 +453,10 @@ static int ocfs2_mknod(struct inode *dir, d_instantiate(dentry, inode); status = 0; leave: + if (default_acl) + posix_acl_release(default_acl); + if (acl) + posix_acl_release(acl); if (status < 0 && did_quota_inode) dquot_free_inode(inode); if (handle) @@ -432,7 +468,6 @@ leave: brelse(new_fe_bh); brelse(parent_fe_bh); - kfree(si.name); kfree(si.value); ocfs2_free_dir_lookup_result(&lookup); @@ -452,6 +487,9 @@ leave: * ocfs2_delete_inode will mutex_lock again. */ if ((status < 0) && inode) { + if (dl) + ocfs2_cleanup_add_entry_failure(osb, dentry, inode); + OCFS2_I(inode)->ip_flags |= OCFS2_INODE_SKIP_ORPHAN_DIR; clear_nlink(inode); iput(inode); @@ -477,6 +515,7 @@ static int __ocfs2_mknod_locked(struct inode *dir, struct ocfs2_dinode *fe = NULL; struct ocfs2_extent_list *fel; u16 feat; + struct ocfs2_inode_info *oi = OCFS2_I(inode); *new_fe_bh = NULL; @@ -491,7 +530,7 @@ static int __ocfs2_mknod_locked(struct inode *dir, *new_fe_bh = sb_getblk(osb->sb, fe_blkno); if (!*new_fe_bh) { - status = -EIO; + status = -ENOMEM; mlog_errno(status); goto leave; } @@ -514,8 +553,8 @@ static int __ocfs2_mknod_locked(struct inode *dir, fe->i_suballoc_loc = cpu_to_le64(suballoc_loc); fe->i_suballoc_bit = cpu_to_le16(suballoc_bit); fe->i_suballoc_slot = cpu_to_le16(inode_ac->ac_alloc_slot); - fe->i_uid = cpu_to_le32(inode->i_uid); - fe->i_gid = cpu_to_le32(inode->i_gid); + fe->i_uid = cpu_to_le32(i_uid_read(inode)); + fe->i_gid = cpu_to_le32(i_gid_read(inode)); fe->i_mode = cpu_to_le16(inode->i_mode); if (S_ISCHR(inode->i_mode) || S_ISBLK(inode->i_mode)) fe->id1.dev1.i_rdev = cpu_to_le64(huge_encode_dev(dev)); @@ -524,7 +563,7 @@ static int __ocfs2_mknod_locked(struct inode *dir, fe->i_last_eb_blk = 0; strcpy(fe->i_signature, OCFS2_INODE_SIGNATURE); - le32_add_cpu(&fe->i_flags, OCFS2_VALID_FL); + fe->i_flags |= cpu_to_le32(OCFS2_VALID_FL); fe->i_atime = fe->i_ctime = fe->i_mtime = cpu_to_le64(CURRENT_TIME.tv_sec); fe->i_mtime_nsec = fe->i_ctime_nsec = fe->i_atime_nsec = @@ -558,8 +597,8 @@ static int __ocfs2_mknod_locked(struct inode *dir, mlog_errno(status); } - status = 0; /* error in ocfs2_create_new_inode_locks is not - * critical */ + oi->i_sync_tid = handle->h_transaction->t_tid; + oi->i_datasync_tid = handle->h_transaction->t_tid; leave: if (status < 0) { @@ -604,7 +643,7 @@ static int ocfs2_mknod_locked(struct ocfs2_super *osb, static int ocfs2_mkdir(struct inode *dir, struct dentry *dentry, - int mode) + umode_t mode) { int ret; @@ -619,8 +658,8 @@ static int ocfs2_mkdir(struct inode *dir, static int ocfs2_create(struct inode *dir, struct dentry *dentry, - int mode, - struct nameidata *nd) + umode_t mode, + bool excl) { int ret; @@ -646,6 +685,7 @@ static int ocfs2_link(struct dentry *old_dentry, struct ocfs2_super *osb = OCFS2_SB(dir->i_sb); struct ocfs2_dir_lookup_result lookup = { NULL, }; sigset_t oldset; + u64 old_de_ino; trace_ocfs2_link((unsigned long long)OCFS2_I(inode)->ip_blkno, old_dentry->d_name.len, old_dentry->d_name.name, @@ -668,6 +708,22 @@ static int ocfs2_link(struct dentry *old_dentry, goto out; } + err = ocfs2_lookup_ino_from_name(dir, old_dentry->d_name.name, + old_dentry->d_name.len, &old_de_ino); + if (err) { + err = -ENOENT; + goto out; + } + + /* + * Check whether another node removed the source inode while we + * were in the vfs. + */ + if (old_de_ino != OCFS2_I(inode)->ip_blkno) { + err = -ENOENT; + goto out; + } + err = ocfs2_check_dir_for_entry(dir, dentry->d_name.name, dentry->d_name.len); if (err) @@ -775,7 +831,7 @@ static int ocfs2_remote_dentry_delete(struct dentry *dentry) return ret; } -static inline int inode_is_unlinkable(struct inode *inode) +static inline int ocfs2_inode_is_unlinkable(struct inode *inode) { if (S_ISDIR(inode->i_mode)) { if (inode->i_nlink == 2) @@ -793,6 +849,7 @@ static int ocfs2_unlink(struct inode *dir, { int status; int child_locked = 0; + bool is_unlinkable = false; struct inode *inode = dentry->d_inode; struct inode *orphan_dir = NULL; struct ocfs2_super *osb = OCFS2_SB(dir->i_sb); @@ -867,7 +924,7 @@ static int ocfs2_unlink(struct inode *dir, goto leave; } - if (inode_is_unlinkable(inode)) { + if (ocfs2_inode_is_unlinkable(inode)) { status = ocfs2_prepare_orphan_dir(osb, &orphan_dir, OCFS2_I(inode)->ip_blkno, orphan_name, &orphan_insert); @@ -875,6 +932,7 @@ static int ocfs2_unlink(struct inode *dir, mlog_errno(status); goto leave; } + is_unlinkable = true; } handle = ocfs2_start_trans(osb, ocfs2_unlink_credits(osb->sb)); @@ -894,15 +952,6 @@ static int ocfs2_unlink(struct inode *dir, fe = (struct ocfs2_dinode *) fe_bh->b_data; - if (inode_is_unlinkable(inode)) { - status = ocfs2_orphan_add(osb, handle, inode, fe_bh, orphan_name, - &orphan_insert, orphan_dir); - if (status < 0) { - mlog_errno(status); - goto leave; - } - } - /* delete the name from the parent dir */ status = ocfs2_delete_entry(handle, dir, &lookup); if (status < 0) { @@ -925,6 +974,14 @@ static int ocfs2_unlink(struct inode *dir, mlog_errno(status); if (S_ISDIR(inode->i_mode)) inc_nlink(dir); + goto leave; + } + + if (is_unlinkable) { + status = ocfs2_orphan_add(osb, handle, inode, fe_bh, + orphan_name, &orphan_insert, orphan_dir); + if (status < 0) + mlog_errno(status); } leave: @@ -949,12 +1006,71 @@ leave: ocfs2_free_dir_lookup_result(&orphan_insert); ocfs2_free_dir_lookup_result(&lookup); - if (status) + if (status && (status != -ENOTEMPTY) && (status != -ENOENT)) mlog_errno(status); return status; } +static int ocfs2_check_if_ancestor(struct ocfs2_super *osb, + u64 src_inode_no, u64 dest_inode_no) +{ + int ret = 0, i = 0; + u64 parent_inode_no = 0; + u64 child_inode_no = src_inode_no; + struct inode *child_inode; + +#define MAX_LOOKUP_TIMES 32 + while (1) { + child_inode = ocfs2_iget(osb, child_inode_no, 0, 0); + if (IS_ERR(child_inode)) { + ret = PTR_ERR(child_inode); + break; + } + + ret = ocfs2_inode_lock(child_inode, NULL, 0); + if (ret < 0) { + iput(child_inode); + if (ret != -ENOENT) + mlog_errno(ret); + break; + } + + ret = ocfs2_lookup_ino_from_name(child_inode, "..", 2, + &parent_inode_no); + ocfs2_inode_unlock(child_inode, 0); + iput(child_inode); + if (ret < 0) { + ret = -ENOENT; + break; + } + + if (parent_inode_no == dest_inode_no) { + ret = 1; + break; + } + + if (parent_inode_no == osb->root_inode->i_ino) { + ret = 0; + break; + } + + child_inode_no = parent_inode_no; + + if (++i >= MAX_LOOKUP_TIMES) { + mlog(ML_NOTICE, "max lookup times reached, filesystem " + "may have nested directories, " + "src inode: %llu, dest inode: %llu.\n", + (unsigned long long)src_inode_no, + (unsigned long long)dest_inode_no); + ret = 0; + break; + } + } + + return ret; +} + /* * The only place this should be used is rename! * if they have the same id, then the 1st one is the only one locked. @@ -966,6 +1082,7 @@ static int ocfs2_double_lock(struct ocfs2_super *osb, struct inode *inode2) { int status; + int inode1_is_ancestor, inode2_is_ancestor; struct ocfs2_inode_info *oi1 = OCFS2_I(inode1); struct ocfs2_inode_info *oi2 = OCFS2_I(inode2); struct buffer_head **tmpbh; @@ -979,9 +1096,26 @@ static int ocfs2_double_lock(struct ocfs2_super *osb, if (*bh2) *bh2 = NULL; - /* we always want to lock the one with the lower lockid first. */ + /* we always want to lock the one with the lower lockid first. + * and if they are nested, we lock ancestor first */ if (oi1->ip_blkno != oi2->ip_blkno) { - if (oi1->ip_blkno < oi2->ip_blkno) { + inode1_is_ancestor = ocfs2_check_if_ancestor(osb, oi2->ip_blkno, + oi1->ip_blkno); + if (inode1_is_ancestor < 0) { + status = inode1_is_ancestor; + goto bail; + } + + inode2_is_ancestor = ocfs2_check_if_ancestor(osb, oi1->ip_blkno, + oi2->ip_blkno); + if (inode2_is_ancestor < 0) { + status = inode2_is_ancestor; + goto bail; + } + + if ((inode1_is_ancestor == 1) || + (oi1->ip_blkno < oi2->ip_blkno && + inode2_is_ancestor == 0)) { /* switch id1 and id2 around */ tmpbh = bh2; bh2 = bh1; @@ -1055,13 +1189,14 @@ static int ocfs2_rename(struct inode *old_dir, handle_t *handle = NULL; struct buffer_head *old_dir_bh = NULL; struct buffer_head *new_dir_bh = NULL; - nlink_t old_dir_nlink = old_dir->i_nlink; + u32 old_dir_nlink = old_dir->i_nlink; struct ocfs2_dinode *old_di; struct ocfs2_dir_lookup_result old_inode_dot_dot_res = { NULL, }; struct ocfs2_dir_lookup_result target_lookup_res = { NULL, }; struct ocfs2_dir_lookup_result old_entry_lookup = { NULL, }; struct ocfs2_dir_lookup_result orphan_insert = { NULL, }; struct ocfs2_dir_lookup_result target_insert = { NULL, }; + bool should_add_orphan = false; /* At some point it might be nice to break this function up a * bit. */ @@ -1098,6 +1233,21 @@ static int ocfs2_rename(struct inode *old_dir, goto bail; } rename_lock = 1; + + /* here we cannot guarantee the inodes haven't just been + * changed, so check if they are nested again */ + status = ocfs2_check_if_ancestor(osb, new_dir->i_ino, + old_inode->i_ino); + if (status < 0) { + mlog_errno(status); + goto bail; + } else if (status == 1) { + status = -EPERM; + trace_ocfs2_rename_not_permitted( + (unsigned long long)old_inode->i_ino, + (unsigned long long)new_dir->i_ino); + goto bail; + } } /* if old and new are the same, this'll just do one lock. */ @@ -1268,6 +1418,7 @@ static int ocfs2_rename(struct inode *old_dir, mlog_errno(status); goto bail; } + should_add_orphan = true; } } else { BUG_ON(new_dentry->d_parent->d_inode != new_dir); @@ -1312,17 +1463,6 @@ static int ocfs2_rename(struct inode *old_dir, goto bail; } - if (S_ISDIR(new_inode->i_mode) || - (ocfs2_read_links_count(newfe) == 1)) { - status = ocfs2_orphan_add(osb, handle, new_inode, - newfe_bh, orphan_name, - &orphan_insert, orphan_dir); - if (status < 0) { - mlog_errno(status); - goto bail; - } - } - /* change the dirent to point to the correct inode */ status = ocfs2_update_entry(new_dir, handle, &target_lookup_res, old_inode); @@ -1337,6 +1477,15 @@ static int ocfs2_rename(struct inode *old_dir, else ocfs2_add_links_count(newfe, -1); ocfs2_journal_dirty(handle, newfe_bh); + if (should_add_orphan) { + status = ocfs2_orphan_add(osb, handle, new_inode, + newfe_bh, orphan_name, + &orphan_insert, orphan_dir); + if (status < 0) { + mlog_errno(status); + goto bail; + } + } } else { /* if the name was not found in new_dir, add it now */ status = ocfs2_add_entry(handle, new_dentry, old_inode, @@ -1379,7 +1528,7 @@ static int ocfs2_rename(struct inode *old_dir, } if (new_inode) { - new_inode->i_nlink--; + drop_nlink(new_inode); new_inode->i_ctime = CURRENT_TIME; } old_dir->i_ctime = old_dir->i_mtime = CURRENT_TIME; @@ -1387,9 +1536,9 @@ static int ocfs2_rename(struct inode *old_dir, if (update_dot_dot) { status = ocfs2_update_entry(old_inode, handle, &old_inode_dot_dot_res, new_dir); - old_dir->i_nlink--; + drop_nlink(old_dir); if (new_inode) { - new_inode->i_nlink--; + drop_nlink(new_inode); } else { inc_nlink(new_dir); mark_inode_dirty(new_dir); @@ -1606,6 +1755,7 @@ static int ocfs2_symlink(struct inode *dir, struct ocfs2_dir_lookup_result lookup = { NULL, }; sigset_t oldset; int did_block_signals = 0; + struct ocfs2_dentry_lock *dl = NULL; trace_ocfs2_symlink_begin(dir, dentry, symname, dentry->d_name.len, dentry->d_name.name); @@ -1726,15 +1876,16 @@ static int ocfs2_symlink(struct inode *dir, fe = (struct ocfs2_dinode *) new_fe_bh->b_data; inode->i_rdev = 0; newsize = l - 1; + inode->i_op = &ocfs2_symlink_inode_operations; if (l > ocfs2_fast_symlink_chars(sb)) { u32 offset = 0; - inode->i_op = &ocfs2_symlink_inode_operations; status = dquot_alloc_space_nodirty(inode, ocfs2_clusters_to_bytes(osb->sb, 1)); if (status) goto bail; did_quota = 1; + inode->i_mapping->a_ops = &ocfs2_aops; status = ocfs2_add_inode_data(osb, inode, &offset, 1, 0, new_fe_bh, handle, data_ac, NULL, @@ -1752,7 +1903,7 @@ static int ocfs2_symlink(struct inode *dir, i_size_write(inode, newsize); inode->i_blocks = ocfs2_inode_sector_count(inode); } else { - inode->i_op = &ocfs2_fast_symlink_inode_operations; + inode->i_mapping->a_ops = &ocfs2_fast_symlink_aops; memcpy((char *) fe->id2.i_symlink, symname, l); i_size_write(inode, newsize); inode->i_blocks = 0; @@ -1793,6 +1944,8 @@ static int ocfs2_symlink(struct inode *dir, goto bail; } + dl = dentry->d_fsdata; + status = ocfs2_add_entry(handle, dentry, inode, le64_to_cpu(fe->i_blkno), parent_fe_bh, &lookup); @@ -1818,7 +1971,6 @@ bail: brelse(new_fe_bh); brelse(parent_fe_bh); - kfree(si.name); kfree(si.value); ocfs2_free_dir_lookup_result(&lookup); if (inode_ac) @@ -1828,6 +1980,9 @@ bail: if (xattr_ac) ocfs2_free_alloc_context(xattr_ac); if ((status < 0) && inode) { + if (dl) + ocfs2_cleanup_add_entry_failure(osb, dentry, inode); + OCFS2_I(inode)->ip_flags |= OCFS2_INODE_SKIP_ORPHAN_DIR; clear_nlink(inode); iput(inode); @@ -2013,23 +2168,6 @@ static int ocfs2_orphan_add(struct ocfs2_super *osb, goto leave; } - /* we're a cluster, and nlink can change on disk from - * underneath us... */ - orphan_fe = (struct ocfs2_dinode *) orphan_dir_bh->b_data; - if (S_ISDIR(inode->i_mode)) - ocfs2_add_links_count(orphan_fe, 1); - orphan_dir_inode->i_nlink = ocfs2_read_links_count(orphan_fe); - ocfs2_journal_dirty(handle, orphan_dir_bh); - - status = __ocfs2_add_entry(handle, orphan_dir_inode, name, - OCFS2_ORPHAN_NAMELEN, inode, - OCFS2_I(inode)->ip_blkno, - orphan_dir_bh, lookup); - if (status < 0) { - mlog_errno(status); - goto leave; - } - /* * We're going to journal the change of i_flags and i_orphaned_slot. * It's safe anyway, though some callers may duplicate the journaling. @@ -2045,7 +2183,24 @@ static int ocfs2_orphan_add(struct ocfs2_super *osb, goto leave; } - le32_add_cpu(&fe->i_flags, OCFS2_ORPHANED_FL); + /* we're a cluster, and nlink can change on disk from + * underneath us... */ + orphan_fe = (struct ocfs2_dinode *) orphan_dir_bh->b_data; + if (S_ISDIR(inode->i_mode)) + ocfs2_add_links_count(orphan_fe, 1); + set_nlink(orphan_dir_inode, ocfs2_read_links_count(orphan_fe)); + ocfs2_journal_dirty(handle, orphan_dir_bh); + + status = __ocfs2_add_entry(handle, orphan_dir_inode, name, + OCFS2_ORPHAN_NAMELEN, inode, + OCFS2_I(inode)->ip_blkno, + orphan_dir_bh, lookup); + if (status < 0) { + mlog_errno(status); + goto rollback; + } + + fe->i_flags |= cpu_to_le32(OCFS2_ORPHANED_FL); OCFS2_I(inode)->ip_flags &= ~OCFS2_INODE_SKIP_ORPHAN_DIR; /* Record which orphan dir our inode now resides @@ -2058,11 +2213,16 @@ static int ocfs2_orphan_add(struct ocfs2_super *osb, trace_ocfs2_orphan_add_end((unsigned long long)OCFS2_I(inode)->ip_blkno, osb->slot_num); +rollback: + if (status < 0) { + if (S_ISDIR(inode->i_mode)) + ocfs2_add_links_count(orphan_fe, -1); + set_nlink(orphan_dir_inode, ocfs2_read_links_count(orphan_fe)); + } + leave: brelse(orphan_dir_bh); - if (status) - mlog_errno(status); return status; } @@ -2116,7 +2276,7 @@ int ocfs2_orphan_del(struct ocfs2_super *osb, orphan_fe = (struct ocfs2_dinode *) orphan_dir_bh->b_data; if (S_ISDIR(inode->i_mode)) ocfs2_add_links_count(orphan_fe, -1); - orphan_dir_inode->i_nlink = ocfs2_read_links_count(orphan_fe); + set_nlink(orphan_dir_inode, ocfs2_read_links_count(orphan_fe)); ocfs2_journal_dirty(handle, orphan_dir_bh); leave: @@ -2217,7 +2377,7 @@ out: brelse(orphan_dir_bh); - return 0; + return ret; } int ocfs2_create_inode_in_orphan(struct inode *dir, @@ -2282,7 +2442,7 @@ int ocfs2_create_inode_in_orphan(struct inode *dir, goto leave; } - inode->i_nlink = 0; + clear_nlink(inode); /* do the real work now. */ status = __ocfs2_mknod_locked(dir, inode, 0, &new_di_bh, parent_di_bh, handle, @@ -2435,10 +2595,11 @@ int ocfs2_mv_orphaned_inode_to_new(struct inode *dir, } di = (struct ocfs2_dinode *)di_bh->b_data; - le32_add_cpu(&di->i_flags, -OCFS2_ORPHANED_FL); + di->i_flags &= ~cpu_to_le32(OCFS2_ORPHANED_FL); di->i_orphaned_slot = 0; - inode->i_nlink = 1; + set_nlink(inode, 1); ocfs2_set_links_count(di, inode->i_nlink); + ocfs2_update_inode_fsync_trans(handle, inode, 1); ocfs2_journal_dirty(handle, di_bh); status = ocfs2_add_entry(handle, dentry, inode, @@ -2498,4 +2659,6 @@ const struct inode_operations ocfs2_dir_iops = { .listxattr = ocfs2_listxattr, .removexattr = generic_removexattr, .fiemap = ocfs2_fiemap, + .get_acl = ocfs2_iop_get_acl, + .set_acl = ocfs2_iop_set_acl, }; diff --git a/fs/ocfs2/ocfs2.h b/fs/ocfs2/ocfs2.h index 409285854f6..bbec539230f 100644 --- a/fs/ocfs2/ocfs2.h +++ b/fs/ocfs2/ocfs2.h @@ -30,6 +30,7 @@ #include <linux/sched.h> #include <linux/wait.h> #include <linux/list.h> +#include <linux/llist.h> #include <linux/rbtree.h> #include <linux/workqueue.h> #include <linux/kref.h> @@ -274,19 +275,16 @@ enum ocfs2_mount_options OCFS2_MOUNT_HB_GLOBAL = 1 << 14, /* Global heartbeat */ }; -#define OCFS2_OSB_SOFT_RO 0x0001 -#define OCFS2_OSB_HARD_RO 0x0002 -#define OCFS2_OSB_ERROR_FS 0x0004 -#define OCFS2_OSB_DROP_DENTRY_LOCK_IMMED 0x0008 - -#define OCFS2_DEFAULT_ATIME_QUANTUM 60 +#define OCFS2_OSB_SOFT_RO 0x0001 +#define OCFS2_OSB_HARD_RO 0x0002 +#define OCFS2_OSB_ERROR_FS 0x0004 +#define OCFS2_DEFAULT_ATIME_QUANTUM 60 struct ocfs2_journal; struct ocfs2_slot_info; struct ocfs2_recovery_map; struct ocfs2_replay_map; struct ocfs2_quota_recovery; -struct ocfs2_dentry_lock; struct ocfs2_super { struct task_struct *commit_task; @@ -347,7 +345,6 @@ struct ocfs2_super struct task_struct *recovery_thread_task; int disable_recovery; wait_queue_head_t checkpoint_event; - atomic_t needs_checkpoint; struct ocfs2_journal *journal; unsigned long osb_commit_interval; @@ -388,6 +385,7 @@ struct ocfs2_super u8 osb_stackflags; char osb_cluster_stack[OCFS2_STACK_LABEL_LEN + 1]; + char osb_cluster_name[OCFS2_CLUSTER_NAME_LEN + 1]; struct ocfs2_cluster_connection *cconn; struct ocfs2_lock_res osb_super_lockres; struct ocfs2_lock_res osb_rename_lockres; @@ -414,10 +412,9 @@ struct ocfs2_super struct list_head blocked_lock_list; unsigned long blocked_lock_count; - /* List of dentry locks to release. Anyone can add locks to - * the list, ocfs2_wq processes the list */ - struct ocfs2_dentry_lock *dentry_lock_list; - struct work_struct dentry_lock_work; + /* List of dquot structures to drop last reference to */ + struct llist_head dquot_drop_list; + struct work_struct dquot_drop_work; wait_queue_head_t osb_mount_event; @@ -425,6 +422,7 @@ struct ocfs2_super struct inode *osb_tl_inode; struct buffer_head *osb_tl_bh; struct delayed_work osb_truncate_log_wq; + atomic_t osb_tl_disable; /* * How many clusters in our truncate log. * It must be protected by osb_tl_inode->i_mutex. @@ -449,6 +447,8 @@ struct ocfs2_super /* rb tree root for refcount lock. */ struct rb_root osb_rf_lock_tree; struct ocfs2_refcount_tree *osb_ref_tree_lru; + + struct mutex system_file_mutex; }; #define OCFS2_SB(sb) ((struct ocfs2_super *)(sb)->s_fs_info) @@ -579,18 +579,6 @@ static inline void ocfs2_set_osb_flag(struct ocfs2_super *osb, spin_unlock(&osb->osb_lock); } - -static inline unsigned long ocfs2_test_osb_flag(struct ocfs2_super *osb, - unsigned long flag) -{ - unsigned long ret; - - spin_lock(&osb->osb_lock); - ret = osb->osb_flags & flag; - spin_unlock(&osb->osb_lock); - return ret; -} - static inline void ocfs2_set_ro_flag(struct ocfs2_super *osb, int hard) { @@ -836,18 +824,65 @@ static inline unsigned int ocfs2_clusters_to_megabytes(struct super_block *sb, static inline void _ocfs2_set_bit(unsigned int bit, unsigned long *bitmap) { - __test_and_set_bit_le(bit, bitmap); + __set_bit_le(bit, bitmap); } #define ocfs2_set_bit(bit, addr) _ocfs2_set_bit((bit), (unsigned long *)(addr)) static inline void _ocfs2_clear_bit(unsigned int bit, unsigned long *bitmap) { - __test_and_clear_bit_le(bit, bitmap); + __clear_bit_le(bit, bitmap); } #define ocfs2_clear_bit(bit, addr) _ocfs2_clear_bit((bit), (unsigned long *)(addr)) #define ocfs2_test_bit test_bit_le #define ocfs2_find_next_zero_bit find_next_zero_bit_le #define ocfs2_find_next_bit find_next_bit_le + +static inline void *correct_addr_and_bit_unaligned(int *bit, void *addr) +{ +#if BITS_PER_LONG == 64 + *bit += ((unsigned long) addr & 7UL) << 3; + addr = (void *) ((unsigned long) addr & ~7UL); +#elif BITS_PER_LONG == 32 + *bit += ((unsigned long) addr & 3UL) << 3; + addr = (void *) ((unsigned long) addr & ~3UL); +#else +#error "how many bits you are?!" +#endif + return addr; +} + +static inline void ocfs2_set_bit_unaligned(int bit, void *bitmap) +{ + bitmap = correct_addr_and_bit_unaligned(&bit, bitmap); + ocfs2_set_bit(bit, bitmap); +} + +static inline void ocfs2_clear_bit_unaligned(int bit, void *bitmap) +{ + bitmap = correct_addr_and_bit_unaligned(&bit, bitmap); + ocfs2_clear_bit(bit, bitmap); +} + +static inline int ocfs2_test_bit_unaligned(int bit, void *bitmap) +{ + bitmap = correct_addr_and_bit_unaligned(&bit, bitmap); + return ocfs2_test_bit(bit, bitmap); +} + +static inline int ocfs2_find_next_zero_bit_unaligned(void *bitmap, int max, + int start) +{ + int fix = 0, ret, tmpmax; + bitmap = correct_addr_and_bit_unaligned(&fix, bitmap); + tmpmax = max + fix; + start += fix; + + ret = ocfs2_find_next_zero_bit(bitmap, tmpmax, start) - fix; + if (ret > max) + return max; + return ret; +} + #endif /* OCFS2_H */ diff --git a/fs/ocfs2/ocfs2_fs.h b/fs/ocfs2/ocfs2_fs.h index b68f87a8392..938387a10d5 100644 --- a/fs/ocfs2/ocfs2_fs.h +++ b/fs/ocfs2/ocfs2_fs.h @@ -1019,7 +1019,7 @@ struct ocfs2_xattr_entry { __le16 xe_name_offset; /* byte offset from the 1st entry in the local xattr storage(inode, xattr block or xattr bucket). */ - __u8 xe_name_len; /* xattr name len, does't include prefix. */ + __u8 xe_name_len; /* xattr name len, doesn't include prefix. */ __u8 xe_type; /* the low 7 bits indicate the name prefix * type and the highest bit indicates whether * the EA is stored in the local storage. */ diff --git a/fs/ocfs2/ocfs2_ioctl.h b/fs/ocfs2/ocfs2_ioctl.h index b46f39bf743..5b27ff1fa57 100644 --- a/fs/ocfs2/ocfs2_ioctl.h +++ b/fs/ocfs2/ocfs2_ioctl.h @@ -142,6 +142,38 @@ struct ocfs2_info_journal_size { __u64 ij_journal_size; }; +struct ocfs2_info_freeinode { + struct ocfs2_info_request ifi_req; + struct ocfs2_info_local_freeinode { + __u64 lfi_total; + __u64 lfi_free; + } ifi_stat[OCFS2_MAX_SLOTS]; + __u32 ifi_slotnum; /* out */ + __u32 ifi_pad; +}; + +#define OCFS2_INFO_MAX_HIST (32) + +struct ocfs2_info_freefrag { + struct ocfs2_info_request iff_req; + struct ocfs2_info_freefrag_stats { /* (out) */ + struct ocfs2_info_free_chunk_list { + __u32 fc_chunks[OCFS2_INFO_MAX_HIST]; + __u32 fc_clusters[OCFS2_INFO_MAX_HIST]; + } ffs_fc_hist; + __u32 ffs_clusters; + __u32 ffs_free_clusters; + __u32 ffs_free_chunks; + __u32 ffs_free_chunks_real; + __u32 ffs_min; /* Minimum free chunksize in clusters */ + __u32 ffs_max; + __u32 ffs_avg; + __u32 ffs_pad; + } iff_ffs; + __u32 iff_chunksize; /* chunksize in clusters(in) */ + __u32 iff_pad; +}; + /* Codes for ocfs2_info_request */ enum ocfs2_info_type { OCFS2_INFO_CLUSTERSIZE = 1, @@ -151,6 +183,8 @@ enum ocfs2_info_type { OCFS2_INFO_UUID, OCFS2_INFO_FS_FEATURES, OCFS2_INFO_JOURNAL_SIZE, + OCFS2_INFO_FREEINODE, + OCFS2_INFO_FREEFRAG, OCFS2_INFO_NUM_TYPES }; @@ -171,4 +205,38 @@ enum ocfs2_info_type { #define OCFS2_IOC_INFO _IOR('o', 5, struct ocfs2_info) +struct ocfs2_move_extents { +/* All values are in bytes */ + /* in */ + __u64 me_start; /* Virtual start in the file to move */ + __u64 me_len; /* Length of the extents to be moved */ + __u64 me_goal; /* Physical offset of the goal, + it's in block unit */ + __u64 me_threshold; /* Maximum distance from goal or threshold + for auto defragmentation */ + __u64 me_flags; /* Flags for the operation: + * - auto defragmentation. + * - refcount,xattr cases. + */ + /* out */ + __u64 me_moved_len; /* Moved/defraged length */ + __u64 me_new_offset; /* Resulting physical location */ + __u32 me_reserved[2]; /* Reserved for futhure */ +}; + +#define OCFS2_MOVE_EXT_FL_AUTO_DEFRAG (0x00000001) /* Kernel manages to + claim new clusters + as the goal place + for extents moving */ +#define OCFS2_MOVE_EXT_FL_PART_DEFRAG (0x00000002) /* Allow partial extent + moving, is to make + movement less likely + to fail, may make fs + even more fragmented */ +#define OCFS2_MOVE_EXT_FL_COMPLETE (0x00000004) /* Move or defragmenation + completely gets done. + */ + +#define OCFS2_IOC_MOVE_EXT _IOW('o', 6, struct ocfs2_move_extents) + #endif /* OCFS2_IOCTL_H */ diff --git a/fs/ocfs2/ocfs2_trace.h b/fs/ocfs2/ocfs2_trace.h index a1dae5bb54a..6cb019b7c6a 100644 --- a/fs/ocfs2/ocfs2_trace.h +++ b/fs/ocfs2/ocfs2_trace.h @@ -688,6 +688,31 @@ TRACE_EVENT(ocfs2_cache_block_dealloc, __entry->blkno, __entry->bit) ); +TRACE_EVENT(ocfs2_trim_extent, + TP_PROTO(struct super_block *sb, unsigned long long blk, + unsigned long long count), + TP_ARGS(sb, blk, count), + TP_STRUCT__entry( + __field(int, dev_major) + __field(int, dev_minor) + __field(unsigned long long, blk) + __field(__u64, count) + ), + TP_fast_assign( + __entry->dev_major = MAJOR(sb->s_dev); + __entry->dev_minor = MINOR(sb->s_dev); + __entry->blk = blk; + __entry->count = count; + ), + TP_printk("%d %d %llu %llu", + __entry->dev_major, __entry->dev_minor, + __entry->blk, __entry->count) +); + +DEFINE_OCFS2_ULL_UINT_UINT_UINT_EVENT(ocfs2_trim_group); + +DEFINE_OCFS2_ULL_ULL_ULL_EVENT(ocfs2_trim_fs); + /* End of trace events for fs/ocfs2/alloc.c. */ /* Trace events for fs/ocfs2/localalloc.c. */ @@ -2267,6 +2292,8 @@ TRACE_EVENT(ocfs2_rename, __entry->new_len, __get_str(new_name)) ); +DEFINE_OCFS2_ULL_ULL_EVENT(ocfs2_rename_not_permitted); + TRACE_EVENT(ocfs2_rename_target_exists, TP_PROTO(int new_len, const char *new_name), TP_ARGS(new_len, new_name), @@ -2554,6 +2581,8 @@ DEFINE_OCFS2_INT_INT_EVENT(ocfs2_extend_trans); DEFINE_OCFS2_INT_EVENT(ocfs2_extend_trans_restart); +DEFINE_OCFS2_INT_INT_EVENT(ocfs2_allocate_extend_trans); + DEFINE_OCFS2_ULL_ULL_UINT_UINT_EVENT(ocfs2_journal_access); DEFINE_OCFS2_ULL_EVENT(ocfs2_journal_dirty); diff --git a/fs/ocfs2/quota.h b/fs/ocfs2/quota.h index d5ab56cbe5c..f266d67df3c 100644 --- a/fs/ocfs2/quota.h +++ b/fs/ocfs2/quota.h @@ -28,6 +28,7 @@ struct ocfs2_dquot { unsigned int dq_use_count; /* Number of nodes having reference to this entry in global quota file */ s64 dq_origspace; /* Last globally synced space usage */ s64 dq_originodes; /* Last globally synced inode usage */ + struct llist_node list; /* Member of list of dquots to drop */ }; /* Description of one chunk to recover in memory */ @@ -110,6 +111,7 @@ int ocfs2_read_quota_phys_block(struct inode *inode, u64 p_block, int ocfs2_create_local_dquot(struct dquot *dquot); int ocfs2_local_release_dquot(handle_t *handle, struct dquot *dquot); int ocfs2_local_write_dquot(struct dquot *dquot); +void ocfs2_drop_dquot_refs(struct work_struct *work); extern const struct dquot_operations ocfs2_quota_operations; extern struct quota_format_type ocfs2_quota_format; diff --git a/fs/ocfs2/quota_global.c b/fs/ocfs2/quota_global.c index 92fcd575775..b990a62cff5 100644 --- a/fs/ocfs2/quota_global.c +++ b/fs/ocfs2/quota_global.c @@ -10,6 +10,7 @@ #include <linux/jiffies.h> #include <linux/writeback.h> #include <linux/workqueue.h> +#include <linux/llist.h> #include <cluster/masklog.h> @@ -95,7 +96,7 @@ static void ocfs2_global_mem2diskdqb(void *dp, struct dquot *dquot) struct ocfs2_global_disk_dqblk *d = dp; struct mem_dqblk *m = &dquot->dq_dqb; - d->dqb_id = cpu_to_le32(dquot->dq_id); + d->dqb_id = cpu_to_le32(from_kqid(&init_user_ns, dquot->dq_id)); d->dqb_use_count = cpu_to_le32(OCFS2_DQUOT(dquot)->dq_use_count); d->dqb_ihardlimit = cpu_to_le64(m->dqb_ihardlimit); d->dqb_isoftlimit = cpu_to_le64(m->dqb_isoftlimit); @@ -112,11 +113,14 @@ static int ocfs2_global_is_id(void *dp, struct dquot *dquot) { struct ocfs2_global_disk_dqblk *d = dp; struct ocfs2_mem_dqinfo *oinfo = - sb_dqinfo(dquot->dq_sb, dquot->dq_type)->dqi_priv; + sb_dqinfo(dquot->dq_sb, dquot->dq_id.type)->dqi_priv; if (qtree_entry_unused(&oinfo->dqi_gi, dp)) return 0; - return le32_to_cpu(d->dqb_id) == dquot->dq_id; + + return qid_eq(make_kqid(&init_user_ns, dquot->dq_id.type, + le32_to_cpu(d->dqb_id)), + dquot->dq_id); } struct qtree_fmt_operations ocfs2_global_ops = { @@ -231,7 +235,7 @@ ssize_t ocfs2_quota_write(struct super_block *sb, int type, len = sb->s_blocksize - OCFS2_QBLK_RESERVED_SPACE - offset; } - if (gqinode->i_size < off + len) { + if (i_size_read(gqinode) < off + len) { loff_t rounded_end = ocfs2_align_bytes_to_blocks(sb, off + len); @@ -399,8 +403,6 @@ int ocfs2_global_read_info(struct super_block *sb, int type) msecs_to_jiffies(oinfo->dqi_syncms)); out_err: - if (status) - mlog_errno(status); return status; out_unlock: ocfs2_unlock_global_qf(oinfo, 0); @@ -477,7 +479,7 @@ int __ocfs2_sync_dquot(struct dquot *dquot, int freeing) { int err, err2; struct super_block *sb = dquot->dq_sb; - int type = dquot->dq_type; + int type = dquot->dq_id.type; struct ocfs2_mem_dqinfo *info = sb_dqinfo(sb, type)->dqi_priv; struct ocfs2_global_disk_dqblk dqblk; s64 spacechange, inodechange; @@ -506,7 +508,8 @@ int __ocfs2_sync_dquot(struct dquot *dquot, int freeing) olditime = dquot->dq_dqb.dqb_itime; oldbtime = dquot->dq_dqb.dqb_btime; ocfs2_global_disk2memdqb(dquot, &dqblk); - trace_ocfs2_sync_dquot(dquot->dq_id, dquot->dq_dqb.dqb_curspace, + trace_ocfs2_sync_dquot(from_kqid(&init_user_ns, dquot->dq_id), + dquot->dq_dqb.dqb_curspace, (long long)spacechange, dquot->dq_dqb.dqb_curinodes, (long long)inodechange); @@ -557,8 +560,8 @@ int __ocfs2_sync_dquot(struct dquot *dquot, int freeing) err = ocfs2_qinfo_lock(info, freeing); if (err < 0) { mlog(ML_ERROR, "Failed to lock quota info, losing quota write" - " (type=%d, id=%u)\n", dquot->dq_type, - (unsigned)dquot->dq_id); + " (type=%d, id=%u)\n", dquot->dq_id.type, + (unsigned)from_kqid(&init_user_ns, dquot->dq_id)); goto out; } if (freeing) @@ -593,9 +596,10 @@ static int ocfs2_sync_dquot_helper(struct dquot *dquot, unsigned long type) struct ocfs2_super *osb = OCFS2_SB(sb); int status = 0; - trace_ocfs2_sync_dquot_helper(dquot->dq_id, dquot->dq_type, + trace_ocfs2_sync_dquot_helper(from_kqid(&init_user_ns, dquot->dq_id), + dquot->dq_id.type, type, sb->s_id); - if (type != dquot->dq_type) + if (type != dquot->dq_id.type) goto out; status = ocfs2_lock_global_qf(oinfo, 1); if (status < 0) @@ -645,7 +649,8 @@ static int ocfs2_write_dquot(struct dquot *dquot) struct ocfs2_super *osb = OCFS2_SB(dquot->dq_sb); int status = 0; - trace_ocfs2_write_dquot(dquot->dq_id, dquot->dq_type); + trace_ocfs2_write_dquot(from_kqid(&init_user_ns, dquot->dq_id), + dquot->dq_id.type); handle = ocfs2_start_trans(osb, OCFS2_QWRITE_CREDITS); if (IS_ERR(handle)) { @@ -675,25 +680,60 @@ static int ocfs2_calc_qdel_credits(struct super_block *sb, int type) OCFS2_INODE_UPDATE_CREDITS; } +void ocfs2_drop_dquot_refs(struct work_struct *work) +{ + struct ocfs2_super *osb = container_of(work, struct ocfs2_super, + dquot_drop_work); + struct llist_node *list; + struct ocfs2_dquot *odquot, *next_odquot; + + list = llist_del_all(&osb->dquot_drop_list); + llist_for_each_entry_safe(odquot, next_odquot, list, list) { + /* Drop the reference we acquired in ocfs2_dquot_release() */ + dqput(&odquot->dq_dquot); + } +} + +/* + * Called when the last reference to dquot is dropped. If we are called from + * downconvert thread, we cannot do all the handling here because grabbing + * quota lock could deadlock (the node holding the quota lock could need some + * other cluster lock to proceed but with blocked downconvert thread we cannot + * release any lock). + */ static int ocfs2_release_dquot(struct dquot *dquot) { handle_t *handle; struct ocfs2_mem_dqinfo *oinfo = - sb_dqinfo(dquot->dq_sb, dquot->dq_type)->dqi_priv; + sb_dqinfo(dquot->dq_sb, dquot->dq_id.type)->dqi_priv; struct ocfs2_super *osb = OCFS2_SB(dquot->dq_sb); int status = 0; - trace_ocfs2_release_dquot(dquot->dq_id, dquot->dq_type); + trace_ocfs2_release_dquot(from_kqid(&init_user_ns, dquot->dq_id), + dquot->dq_id.type); mutex_lock(&dquot->dq_lock); /* Check whether we are not racing with some other dqget() */ if (atomic_read(&dquot->dq_count) > 1) goto out; + /* Running from downconvert thread? Postpone quota processing to wq */ + if (current == osb->dc_task) { + /* + * Grab our own reference to dquot and queue it for delayed + * dropping. Quota code rechecks after calling + * ->release_dquot() and won't free dquot structure. + */ + dqgrab(dquot); + /* First entry on list -> queue work */ + if (llist_add(&OCFS2_DQUOT(dquot)->list, &osb->dquot_drop_list)) + queue_work(ocfs2_wq, &osb->dquot_drop_work); + goto out; + } status = ocfs2_lock_global_qf(oinfo, 1); if (status < 0) goto out; handle = ocfs2_start_trans(osb, - ocfs2_calc_qdel_credits(dquot->dq_sb, dquot->dq_type)); + ocfs2_calc_qdel_credits(dquot->dq_sb, dquot->dq_id.type)); if (IS_ERR(handle)) { status = PTR_ERR(handle); mlog_errno(status); @@ -712,6 +752,12 @@ static int ocfs2_release_dquot(struct dquot *dquot) */ if (status < 0) mlog_errno(status); + /* + * Clear dq_off so that we search for the structure in quota file next + * time we acquire it. The structure might be deleted and reallocated + * elsewhere by another node while our dquot structure is on freelist. + */ + dquot->dq_off = 0; clear_bit(DQ_ACTIVE_B, &dquot->dq_flags); out_trans: ocfs2_commit_trans(osb, handle); @@ -735,13 +781,14 @@ static int ocfs2_acquire_dquot(struct dquot *dquot) int ex = 0; struct super_block *sb = dquot->dq_sb; struct ocfs2_super *osb = OCFS2_SB(sb); - int type = dquot->dq_type; + int type = dquot->dq_id.type; struct ocfs2_mem_dqinfo *info = sb_dqinfo(sb, type)->dqi_priv; struct inode *gqinode = info->dqi_gqinode; int need_alloc = ocfs2_global_qinit_alloc(sb, type); handle_t *handle; - trace_ocfs2_acquire_dquot(dquot->dq_id, type); + trace_ocfs2_acquire_dquot(from_kqid(&init_user_ns, dquot->dq_id), + type); mutex_lock(&dquot->dq_lock); /* * We need an exclusive lock, because we're going to update use count @@ -750,16 +797,17 @@ static int ocfs2_acquire_dquot(struct dquot *dquot) status = ocfs2_lock_global_qf(info, 1); if (status < 0) goto out; - if (!test_bit(DQ_READ_B, &dquot->dq_flags)) { - status = ocfs2_qinfo_lock(info, 0); - if (status < 0) - goto out_dq; - status = qtree_read_dquot(&info->dqi_gi, dquot); - ocfs2_qinfo_unlock(info, 0); - if (status < 0) - goto out_dq; - } - set_bit(DQ_READ_B, &dquot->dq_flags); + status = ocfs2_qinfo_lock(info, 0); + if (status < 0) + goto out_dq; + /* + * We always want to read dquot structure from disk because we don't + * know what happened with it while it was on freelist. + */ + status = qtree_read_dquot(&info->dqi_gi, dquot); + ocfs2_qinfo_unlock(info, 0); + if (status < 0) + goto out_dq; OCFS2_DQUOT(dquot)->dq_use_count++; OCFS2_DQUOT(dquot)->dq_origspace = dquot->dq_dqb.dqb_curspace; @@ -772,8 +820,8 @@ static int ocfs2_acquire_dquot(struct dquot *dquot) */ WARN_ON(journal_current_handle()); status = ocfs2_extend_no_holes(gqinode, NULL, - gqinode->i_size + (need_alloc << sb->s_blocksize_bits), - gqinode->i_size); + i_size_read(gqinode) + (need_alloc << sb->s_blocksize_bits), + i_size_read(gqinode)); if (status < 0) goto out_dq; } @@ -823,12 +871,13 @@ static int ocfs2_mark_dquot_dirty(struct dquot *dquot) int sync = 0; int status; struct super_block *sb = dquot->dq_sb; - int type = dquot->dq_type; + int type = dquot->dq_id.type; struct ocfs2_mem_dqinfo *oinfo = sb_dqinfo(sb, type)->dqi_priv; handle_t *handle; struct ocfs2_super *osb = OCFS2_SB(sb); - trace_ocfs2_mark_dquot_dirty(dquot->dq_id, type); + trace_ocfs2_mark_dquot_dirty(from_kqid(&init_user_ns, dquot->dq_id), + type); /* In case user set some limits, sync dquot immediately to global * quota file so that information propagates quicker */ diff --git a/fs/ocfs2/quota_local.c b/fs/ocfs2/quota_local.c index dc8007fc924..2001862bf2b 100644 --- a/fs/ocfs2/quota_local.c +++ b/fs/ocfs2/quota_local.c @@ -404,7 +404,9 @@ struct ocfs2_quota_recovery *ocfs2_begin_quota_recovery( int status = 0; struct ocfs2_quota_recovery *rec; - mlog(ML_NOTICE, "Beginning quota recovery in slot %u\n", slot_num); + printk(KERN_NOTICE "ocfs2: Beginning quota recovery on device (%s) for " + "slot %u\n", osb->dev_str, slot_num); + rec = ocfs2_alloc_quota_recovery(); if (!rec) return ERR_PTR(-ENOMEM); @@ -499,7 +501,9 @@ static int ocfs2_recover_local_quota_file(struct inode *lqinode, } dqblk = (struct ocfs2_local_disk_dqblk *)(qbh->b_data + ol_dqblk_block_off(sb, chunk, bit)); - dquot = dqget(sb, le64_to_cpu(dqblk->dqb_id), type); + dquot = dqget(sb, + make_kqid(&init_user_ns, type, + le64_to_cpu(dqblk->dqb_id))); if (!dquot) { status = -EIO; mlog(ML_ERROR, "Failed to get quota structure " @@ -549,8 +553,8 @@ static int ocfs2_recover_local_quota_file(struct inode *lqinode, goto out_commit; } lock_buffer(qbh); - WARN_ON(!ocfs2_test_bit(bit, dchunk->dqc_bitmap)); - ocfs2_clear_bit(bit, dchunk->dqc_bitmap); + WARN_ON(!ocfs2_test_bit_unaligned(bit, dchunk->dqc_bitmap)); + ocfs2_clear_bit_unaligned(bit, dchunk->dqc_bitmap); le32_add_cpu(&dchunk->dqc_free, 1); unlock_buffer(qbh); ocfs2_journal_dirty(handle, qbh); @@ -596,7 +600,9 @@ int ocfs2_finish_quota_recovery(struct ocfs2_super *osb, struct inode *lqinode; unsigned int flags; - mlog(ML_NOTICE, "Finishing quota recovery in slot %u\n", slot_num); + printk(KERN_NOTICE "ocfs2: Finishing quota recovery on device (%s) for " + "slot %u\n", osb->dev_str, slot_num); + mutex_lock(&sb_dqopt(sb)->dqonoff_mutex); for (type = 0; type < MAXQUOTAS; type++) { if (list_empty(&(rec->r_list[type]))) @@ -612,8 +618,9 @@ int ocfs2_finish_quota_recovery(struct ocfs2_super *osb, /* Someone else is holding the lock? Then he must be * doing the recovery. Just skip the file... */ if (status == -EAGAIN) { - mlog(ML_NOTICE, "skipping quota recovery for slot %d " - "because quota file is locked.\n", slot_num); + printk(KERN_NOTICE "ocfs2: Skipping quota recovery on " + "device (%s) for slot %d because quota file is " + "locked.\n", osb->dev_str, slot_num); status = 0; goto out_put; } else if (status < 0) { @@ -876,7 +883,8 @@ static void olq_set_dquot(struct buffer_head *bh, void *private) dqblk = (struct ocfs2_local_disk_dqblk *)(bh->b_data + ol_dqblk_block_offset(sb, od->dq_local_off)); - dqblk->dqb_id = cpu_to_le64(od->dq_dquot.dq_id); + dqblk->dqb_id = cpu_to_le64(from_kqid(&init_user_ns, + od->dq_dquot.dq_id)); spin_lock(&dq_data_lock); dqblk->dqb_spacemod = cpu_to_le64(od->dq_dquot.dq_dqb.dqb_curspace - od->dq_origspace); @@ -886,7 +894,7 @@ static void olq_set_dquot(struct buffer_head *bh, void *private) trace_olq_set_dquot( (unsigned long long)le64_to_cpu(dqblk->dqb_spacemod), (unsigned long long)le64_to_cpu(dqblk->dqb_inodemod), - od->dq_dquot.dq_id); + from_kqid(&init_user_ns, od->dq_dquot.dq_id)); } /* Write dquot to local quota file */ @@ -895,7 +903,7 @@ int ocfs2_local_write_dquot(struct dquot *dquot) struct super_block *sb = dquot->dq_sb; struct ocfs2_dquot *od = OCFS2_DQUOT(dquot); struct buffer_head *bh; - struct inode *lqinode = sb_dqopt(sb)->files[dquot->dq_type]; + struct inode *lqinode = sb_dqopt(sb)->files[dquot->dq_id.type]; int status; status = ocfs2_read_quota_phys_block(lqinode, od->dq_local_phys_blk, @@ -944,7 +952,7 @@ static struct ocfs2_quota_chunk *ocfs2_find_free_entry(struct super_block *sb, * ol_quota_entries_per_block(sb); } - found = ocfs2_find_next_zero_bit(dchunk->dqc_bitmap, len, 0); + found = ocfs2_find_next_zero_bit_unaligned(dchunk->dqc_bitmap, len, 0); /* We failed? */ if (found == len) { mlog(ML_ERROR, "Did not find empty entry in chunk %d with %u" @@ -974,14 +982,14 @@ static struct ocfs2_quota_chunk *ocfs2_local_quota_add_chunk( /* We are protected by dqio_sem so no locking needed */ status = ocfs2_extend_no_holes(lqinode, NULL, - lqinode->i_size + 2 * sb->s_blocksize, - lqinode->i_size); + i_size_read(lqinode) + 2 * sb->s_blocksize, + i_size_read(lqinode)); if (status < 0) { mlog_errno(status); goto out; } status = ocfs2_simple_size_update(lqinode, oinfo->dqi_lqi_bh, - lqinode->i_size + 2 * sb->s_blocksize); + i_size_read(lqinode) + 2 * sb->s_blocksize); if (status < 0) { mlog_errno(status); goto out; @@ -1117,14 +1125,14 @@ static struct ocfs2_quota_chunk *ocfs2_extend_local_quota_file( /* We are protected by dqio_sem so no locking needed */ status = ocfs2_extend_no_holes(lqinode, NULL, - lqinode->i_size + sb->s_blocksize, - lqinode->i_size); + i_size_read(lqinode) + sb->s_blocksize, + i_size_read(lqinode)); if (status < 0) { mlog_errno(status); goto out; } status = ocfs2_simple_size_update(lqinode, oinfo->dqi_lqi_bh, - lqinode->i_size + sb->s_blocksize); + i_size_read(lqinode) + sb->s_blocksize); if (status < 0) { mlog_errno(status); goto out; @@ -1208,7 +1216,7 @@ static void olq_alloc_dquot(struct buffer_head *bh, void *private) struct ocfs2_local_disk_chunk *dchunk; dchunk = (struct ocfs2_local_disk_chunk *)bh->b_data; - ocfs2_set_bit(*offset, dchunk->dqc_bitmap); + ocfs2_set_bit_unaligned(*offset, dchunk->dqc_bitmap); le32_add_cpu(&dchunk->dqc_free, -1); } @@ -1216,7 +1224,7 @@ static void olq_alloc_dquot(struct buffer_head *bh, void *private) int ocfs2_create_local_dquot(struct dquot *dquot) { struct super_block *sb = dquot->dq_sb; - int type = dquot->dq_type; + int type = dquot->dq_id.type; struct inode *lqinode = sb_dqopt(sb)->files[type]; struct ocfs2_quota_chunk *chunk; struct ocfs2_dquot *od = OCFS2_DQUOT(dquot); @@ -1270,7 +1278,7 @@ out: int ocfs2_local_release_dquot(handle_t *handle, struct dquot *dquot) { int status; - int type = dquot->dq_type; + int type = dquot->dq_id.type; struct ocfs2_dquot *od = OCFS2_DQUOT(dquot); struct super_block *sb = dquot->dq_sb; struct ocfs2_local_disk_chunk *dchunk; @@ -1289,16 +1297,12 @@ int ocfs2_local_release_dquot(handle_t *handle, struct dquot *dquot) (od->dq_chunk->qc_headerbh->b_data); /* Mark structure as freed */ lock_buffer(od->dq_chunk->qc_headerbh); - ocfs2_clear_bit(offset, dchunk->dqc_bitmap); + ocfs2_clear_bit_unaligned(offset, dchunk->dqc_bitmap); le32_add_cpu(&dchunk->dqc_free, 1); unlock_buffer(od->dq_chunk->qc_headerbh); ocfs2_journal_dirty(handle, od->dq_chunk->qc_headerbh); out: - /* Clear the read bit so that next time someone uses this - * dquot he reads fresh info from disk and allocates local - * dquot structure */ - clear_bit(DQ_READ_B, &dquot->dq_flags); return status; } diff --git a/fs/ocfs2/refcounttree.c b/fs/ocfs2/refcounttree.c index 5d32749c896..636aab69ead 100644 --- a/fs/ocfs2/refcounttree.c +++ b/fs/ocfs2/refcounttree.c @@ -46,10 +46,10 @@ #include <linux/quotaops.h> #include <linux/namei.h> #include <linux/mount.h> +#include <linux/posix_acl.h> struct ocfs2_cow_context { struct inode *inode; - struct file *file; u32 cow_start; u32 cow_len; struct ocfs2_extent_tree data_et; @@ -66,7 +66,7 @@ struct ocfs2_cow_context { u32 *num_clusters, unsigned int *extent_flags); int (*cow_duplicate_clusters)(handle_t *handle, - struct ocfs2_cow_context *context, + struct inode *inode, u32 cpos, u32 old_cluster, u32 new_cluster, u32 new_len); }; @@ -613,6 +613,11 @@ static int ocfs2_create_refcount_tree(struct inode *inode, } new_bh = sb_getblk(inode->i_sb, first_blkno); + if (!new_bh) { + ret = -ENOMEM; + mlog_errno(ret); + goto out_commit; + } ocfs2_set_new_buffer_uptodate(&new_tree->rf_ci, new_bh); ret = ocfs2_journal_access_rb(handle, &new_tree->rf_ci, new_bh, @@ -1036,14 +1041,14 @@ static int ocfs2_get_refcount_cpos_end(struct ocfs2_caching_info *ci, tmp_el = left_path->p_node[subtree_root].el; blkno = left_path->p_node[subtree_root+1].bh->b_blocknr; - for (i = 0; i < le32_to_cpu(tmp_el->l_next_free_rec); i++) { + for (i = 0; i < le16_to_cpu(tmp_el->l_next_free_rec); i++) { if (le64_to_cpu(tmp_el->l_recs[i].e_blkno) == blkno) { *cpos_end = le32_to_cpu(tmp_el->l_recs[i+1].e_cpos); break; } } - BUG_ON(i == le32_to_cpu(tmp_el->l_next_free_rec)); + BUG_ON(i == le16_to_cpu(tmp_el->l_next_free_rec)); out: ocfs2_free_path(left_path); @@ -1311,7 +1316,7 @@ static int ocfs2_expand_inline_ref_root(handle_t *handle, new_bh = sb_getblk(sb, blkno); if (new_bh == NULL) { - ret = -EIO; + ret = -ENOMEM; mlog_errno(ret); goto out; } @@ -1403,10 +1408,9 @@ static void swap_refcount_rec(void *a, void *b, int size) { struct ocfs2_refcount_rec *l = a, *r = b, tmp; - tmp = *(struct ocfs2_refcount_rec *)l; - *(struct ocfs2_refcount_rec *)l = - *(struct ocfs2_refcount_rec *)r; - *(struct ocfs2_refcount_rec *)r = tmp; + tmp = *l; + *l = *r; + *r = tmp; } /* @@ -1468,7 +1472,7 @@ static int ocfs2_divide_leaf_refcount_block(struct buffer_head *ref_leaf_bh, trace_ocfs2_divide_leaf_refcount_block( (unsigned long long)ref_leaf_bh->b_blocknr, - le32_to_cpu(rl->rl_count), le32_to_cpu(rl->rl_used)); + le16_to_cpu(rl->rl_count), le16_to_cpu(rl->rl_used)); /* * XXX: Improvement later. @@ -1562,7 +1566,7 @@ static int ocfs2_new_leaf_refcount_block(handle_t *handle, new_bh = sb_getblk(sb, blkno); if (new_bh == NULL) { - ret = -EIO; + ret = -ENOMEM; mlog_errno(ret); goto out; } @@ -2411,7 +2415,7 @@ static int ocfs2_calc_refcount_meta_credits(struct super_block *sb, rb = (struct ocfs2_refcount_block *) prev_bh->b_data; - if (le64_to_cpu(rb->rf_records.rl_used) + + if (le16_to_cpu(rb->rf_records.rl_used) + recs_add > le16_to_cpu(rb->rf_records.rl_count)) ref_blocks++; @@ -2476,7 +2480,7 @@ static int ocfs2_calc_refcount_meta_credits(struct super_block *sb, if (prev_bh) { rb = (struct ocfs2_refcount_block *)prev_bh->b_data; - if (le64_to_cpu(rb->rf_records.rl_used) + recs_add > + if (le16_to_cpu(rb->rf_records.rl_used) + recs_add > le16_to_cpu(rb->rf_records.rl_count)) ref_blocks++; @@ -2503,8 +2507,7 @@ static int ocfs2_calc_refcount_meta_credits(struct super_block *sb, ocfs2_init_refcount_extent_tree(&et, ci, ref_root_bh); *meta_add += ocfs2_extend_meta_needed(et.et_root_el); *credits += ocfs2_calc_extend_credits(sb, - et.et_root_el, - ref_blocks); + et.et_root_el); } else { *credits += OCFS2_EXPAND_REFCOUNT_TREE_CREDITS; *meta_add += 1; @@ -2875,8 +2878,7 @@ static int ocfs2_lock_refcount_allocators(struct super_block *sb, meta_add = ocfs2_extend_meta_needed(et->et_root_el); - *credits += ocfs2_calc_extend_credits(sb, et->et_root_el, - num_clusters + 2); + *credits += ocfs2_calc_extend_credits(sb, et->et_root_el); ret = ocfs2_calc_refcount_meta_credits(sb, ref_ci, ref_root_bh, p_cluster, num_clusters, @@ -2921,20 +2923,19 @@ static int ocfs2_clear_cow_buffer(handle_t *handle, struct buffer_head *bh) return 0; } -static int ocfs2_duplicate_clusters_by_page(handle_t *handle, - struct ocfs2_cow_context *context, - u32 cpos, u32 old_cluster, - u32 new_cluster, u32 new_len) +int ocfs2_duplicate_clusters_by_page(handle_t *handle, + struct inode *inode, + u32 cpos, u32 old_cluster, + u32 new_cluster, u32 new_len) { int ret = 0, partial; - struct ocfs2_caching_info *ci = context->data_et.et_ci; - struct super_block *sb = ocfs2_metadata_cache_get_super(ci); + struct super_block *sb = inode->i_sb; u64 new_block = ocfs2_clusters_to_blocks(sb, new_cluster); struct page *page; pgoff_t page_index; unsigned int from, to, readahead_pages; loff_t offset, end, map_end; - struct address_space *mapping = context->inode->i_mapping; + struct address_space *mapping = inode->i_mapping; trace_ocfs2_duplicate_clusters_by_page(cpos, old_cluster, new_cluster, new_len); @@ -2948,8 +2949,8 @@ static int ocfs2_duplicate_clusters_by_page(handle_t *handle, * We only duplicate pages until we reach the page contains i_size - 1. * So trim 'end' to i_size. */ - if (end > i_size_read(context->inode)) - end = i_size_read(context->inode); + if (end > i_size_read(inode)) + end = i_size_read(inode); while (offset < end) { page_index = offset >> PAGE_CACHE_SHIFT; @@ -2964,6 +2965,11 @@ static int ocfs2_duplicate_clusters_by_page(handle_t *handle, to = map_end & (PAGE_CACHE_SIZE - 1); page = find_or_create_page(mapping, page_index, GFP_NOFS); + if (!page) { + ret = -ENOMEM; + mlog_errno(ret); + break; + } /* * In case PAGE_CACHE_SIZE <= CLUSTER_SIZE, This page @@ -2972,14 +2978,6 @@ static int ocfs2_duplicate_clusters_by_page(handle_t *handle, if (PAGE_CACHE_SIZE <= OCFS2_SB(sb)->s_clustersize) BUG_ON(PageDirty(page)); - if (PageReadahead(page) && context->file) { - page_cache_async_readahead(mapping, - &context->file->f_ra, - context->file, - page, page_index, - readahead_pages); - } - if (!PageUptodate(page)) { ret = block_read_full_page(page, ocfs2_get_block); if (ret) { @@ -2999,7 +2997,7 @@ static int ocfs2_duplicate_clusters_by_page(handle_t *handle, } } - ocfs2_map_and_dirty_page(context->inode, + ocfs2_map_and_dirty_page(inode, handle, from, to, page, 0, &new_block); mark_page_accessed(page); @@ -3015,14 +3013,14 @@ unlock: return ret; } -static int ocfs2_duplicate_clusters_by_jbd(handle_t *handle, - struct ocfs2_cow_context *context, - u32 cpos, u32 old_cluster, - u32 new_cluster, u32 new_len) +int ocfs2_duplicate_clusters_by_jbd(handle_t *handle, + struct inode *inode, + u32 cpos, u32 old_cluster, + u32 new_cluster, u32 new_len) { int ret = 0; - struct super_block *sb = context->inode->i_sb; - struct ocfs2_caching_info *ci = context->data_et.et_ci; + struct super_block *sb = inode->i_sb; + struct ocfs2_caching_info *ci = INODE_CACHE(inode); int i, blocks = ocfs2_clusters_to_blocks(sb, new_len); u64 old_block = ocfs2_clusters_to_blocks(sb, old_cluster); u64 new_block = ocfs2_clusters_to_blocks(sb, new_cluster); @@ -3036,7 +3034,7 @@ static int ocfs2_duplicate_clusters_by_jbd(handle_t *handle, for (i = 0; i < blocks; i++, old_block++, new_block++) { new_bh = sb_getblk(osb->sb, new_block); if (new_bh == NULL) { - ret = -EIO; + ret = -ENOMEM; mlog_errno(ret); break; } @@ -3145,8 +3143,8 @@ static int ocfs2_replace_clusters(handle_t *handle, /*If the old clusters is unwritten, no need to duplicate. */ if (!(ext_flags & OCFS2_EXT_UNWRITTEN)) { - ret = context->cow_duplicate_clusters(handle, context, cpos, - old, new, len); + ret = context->cow_duplicate_clusters(handle, context->inode, + cpos, old, new, len); if (ret) { mlog_errno(ret); goto out; @@ -3162,22 +3160,22 @@ out: return ret; } -static int ocfs2_cow_sync_writeback(struct super_block *sb, - struct ocfs2_cow_context *context, - u32 cpos, u32 num_clusters) +int ocfs2_cow_sync_writeback(struct super_block *sb, + struct inode *inode, + u32 cpos, u32 num_clusters) { int ret = 0; loff_t offset, end, map_end; pgoff_t page_index; struct page *page; - if (ocfs2_should_order_data(context->inode)) + if (ocfs2_should_order_data(inode)) return 0; offset = ((loff_t)cpos) << OCFS2_SB(sb)->s_clustersize_bits; end = offset + (num_clusters << OCFS2_SB(sb)->s_clustersize_bits); - ret = filemap_fdatawrite_range(context->inode->i_mapping, + ret = filemap_fdatawrite_range(inode->i_mapping, offset, end - 1); if (ret < 0) { mlog_errno(ret); @@ -3190,7 +3188,7 @@ static int ocfs2_cow_sync_writeback(struct super_block *sb, if (map_end > end) map_end = end; - page = find_or_create_page(context->inode->i_mapping, + page = find_or_create_page(inode->i_mapping, page_index, GFP_NOFS); BUG_ON(!page); @@ -3349,7 +3347,7 @@ static int ocfs2_make_clusters_writable(struct super_block *sb, * in write-back mode. */ if (context->get_clusters == ocfs2_di_get_clusters) { - ret = ocfs2_cow_sync_writeback(sb, context, cpos, + ret = ocfs2_cow_sync_writeback(sb, context->inode, cpos, orig_num_clusters); if (ret) mlog_errno(ret); @@ -3423,35 +3421,12 @@ static int ocfs2_replace_cow(struct ocfs2_cow_context *context) return ret; } -static void ocfs2_readahead_for_cow(struct inode *inode, - struct file *file, - u32 start, u32 len) -{ - struct address_space *mapping; - pgoff_t index; - unsigned long num_pages; - int cs_bits = OCFS2_SB(inode->i_sb)->s_clustersize_bits; - - if (!file) - return; - - mapping = file->f_mapping; - num_pages = (len << cs_bits) >> PAGE_CACHE_SHIFT; - if (!num_pages) - num_pages = 1; - - index = ((loff_t)start << cs_bits) >> PAGE_CACHE_SHIFT; - page_cache_sync_readahead(mapping, &file->f_ra, file, - index, num_pages); -} - /* * Starting at cpos, try to CoW write_len clusters. Don't CoW * past max_cpos. This will stop when it runs into a hole or an * unrefcounted extent. */ static int ocfs2_refcount_cow_hunk(struct inode *inode, - struct file *file, struct buffer_head *di_bh, u32 cpos, u32 write_len, u32 max_cpos) { @@ -3480,8 +3455,6 @@ static int ocfs2_refcount_cow_hunk(struct inode *inode, BUG_ON(cow_len == 0); - ocfs2_readahead_for_cow(inode, file, cow_start, cow_len); - context = kzalloc(sizeof(struct ocfs2_cow_context), GFP_NOFS); if (!context) { ret = -ENOMEM; @@ -3503,7 +3476,6 @@ static int ocfs2_refcount_cow_hunk(struct inode *inode, context->ref_root_bh = ref_root_bh; context->cow_duplicate_clusters = ocfs2_duplicate_clusters_by_page; context->get_clusters = ocfs2_di_get_clusters; - context->file = file; ocfs2_init_dinode_extent_tree(&context->data_et, INODE_CACHE(inode), di_bh); @@ -3532,7 +3504,6 @@ out: * clusters between cpos and cpos+write_len are safe to modify. */ int ocfs2_refcount_cow(struct inode *inode, - struct file *file, struct buffer_head *di_bh, u32 cpos, u32 write_len, u32 max_cpos) { @@ -3552,7 +3523,7 @@ int ocfs2_refcount_cow(struct inode *inode, num_clusters = write_len; if (ext_flags & OCFS2_EXT_REFCOUNTED) { - ret = ocfs2_refcount_cow_hunk(inode, file, di_bh, cpos, + ret = ocfs2_refcount_cow_hunk(inode, di_bh, cpos, num_clusters, max_cpos); if (ret) { mlog_errno(ret); @@ -3629,7 +3600,7 @@ int ocfs2_refcounted_xattr_delete_need(struct inode *inode, * one will split a refcount rec, so totally we need * clusters * 2 new refcount rec. */ - if (le64_to_cpu(rb->rf_records.rl_used) + clusters * 2 > + if (le16_to_cpu(rb->rf_records.rl_used) + clusters * 2 > le16_to_cpu(rb->rf_records.rl_count)) ref_blocks++; @@ -3657,8 +3628,7 @@ int ocfs2_refcounted_xattr_delete_need(struct inode *inode, ocfs2_init_refcount_extent_tree(&et, ref_ci, ref_root_bh); *credits += ocfs2_calc_extend_credits(inode->i_sb, - et.et_root_el, - ref_blocks); + et.et_root_el); } out: @@ -3706,7 +3676,7 @@ int ocfs2_refcount_cow_xattr(struct inode *inode, context->cow_start = cow_start; context->cow_len = cow_len; context->ref_tree = ref_tree; - context->ref_root_bh = ref_root_bh;; + context->ref_root_bh = ref_root_bh; context->cow_object = xv; context->cow_duplicate_clusters = ocfs2_duplicate_clusters_by_jbd; @@ -3886,7 +3856,10 @@ static int ocfs2_attach_refcount_tree(struct inode *inode, while (cpos < clusters) { ret = ocfs2_get_clusters(inode, cpos, &p_cluster, &num_clusters, &ext_flags); - + if (ret) { + mlog_errno(ret); + goto unlock; + } if (p_cluster && !(ext_flags & OCFS2_EXT_REFCOUNTED)) { ret = ocfs2_add_refcount_flag(inode, &di_et, &ref_tree->rf_ci, @@ -4057,7 +4030,10 @@ static int ocfs2_duplicate_extent_list(struct inode *s_inode, while (cpos < clusters) { ret = ocfs2_get_clusters(s_inode, cpos, &p_cluster, &num_clusters, &ext_flags); - + if (ret) { + mlog_errno(ret); + goto out; + } if (p_cluster) { ret = ocfs2_add_refcounted_extent(t_inode, &et, ref_ci, ref_root_bh, @@ -4292,20 +4268,36 @@ static int ocfs2_reflink(struct dentry *old_dentry, struct inode *dir, struct inode *inode = old_dentry->d_inode; struct buffer_head *old_bh = NULL; struct inode *new_orphan_inode = NULL; + struct posix_acl *default_acl, *acl; + umode_t mode; if (!ocfs2_refcount_tree(OCFS2_SB(inode->i_sb))) return -EOPNOTSUPP; - error = ocfs2_create_inode_in_orphan(dir, inode->i_mode, + mode = inode->i_mode; + error = posix_acl_create(dir, &mode, &default_acl, &acl); + if (error) { + mlog_errno(error); + goto out; + } + + error = ocfs2_create_inode_in_orphan(dir, mode, &new_orphan_inode); if (error) { mlog_errno(error); goto out; } + error = ocfs2_rw_lock(inode, 1); + if (error) { + mlog_errno(error); + goto out; + } + error = ocfs2_inode_lock(inode, &old_bh, 1); if (error) { mlog_errno(error); + ocfs2_rw_unlock(inode, 1); goto out; } @@ -4317,6 +4309,7 @@ static int ocfs2_reflink(struct dentry *old_dentry, struct inode *dir, up_write(&OCFS2_I(inode)->ip_xattr_sem); ocfs2_inode_unlock(inode, 1); + ocfs2_rw_unlock(inode, 1); brelse(old_bh); if (error) { @@ -4327,11 +4320,16 @@ static int ocfs2_reflink(struct dentry *old_dentry, struct inode *dir, /* If the security isn't preserved, we need to re-initialize them. */ if (!preserve) { error = ocfs2_init_security_and_acl(dir, new_orphan_inode, - &new_dentry->d_name); + &new_dentry->d_name, + default_acl, acl); if (error) mlog_errno(error); } out: + if (default_acl) + posix_acl_release(default_acl); + if (acl) + posix_acl_release(acl); if (!error) { error = ocfs2_mv_orphaned_inode_to_new(dir, new_orphan_inode, new_dentry); @@ -4368,25 +4366,6 @@ static inline int ocfs2_may_create(struct inode *dir, struct dentry *child) return inode_permission(dir, MAY_WRITE | MAY_EXEC); } -/* copied from user_path_parent. */ -static int ocfs2_user_path_parent(const char __user *path, - struct nameidata *nd, char **name) -{ - char *s = getname(path); - int error; - - if (IS_ERR(s)) - return PTR_ERR(s); - - error = kern_path_parent(s, nd); - if (error) - putname(s); - else - *name = s; - - return error; -} - /** * ocfs2_vfs_reflink - Create a reference-counted link * @@ -4426,7 +4405,7 @@ static int ocfs2_vfs_reflink(struct dentry *old_dentry, struct inode *dir, * rights to do so. */ if (preserve) { - if ((current_fsuid() != inode->i_uid) && !capable(CAP_CHOWN)) + if (!uid_eq(current_fsuid(), inode->i_uid) && !capable(CAP_CHOWN)) return -EPERM; if (!in_group_p(inode->i_gid) && !capable(CAP_CHOWN)) return -EPERM; @@ -4460,10 +4439,8 @@ int ocfs2_reflink_ioctl(struct inode *inode, bool preserve) { struct dentry *new_dentry; - struct nameidata nd; - struct path old_path; + struct path old_path, new_path; int error; - char *to = NULL; if (!ocfs2_refcount_tree(OCFS2_SB(inode->i_sb))) return -EOPNOTSUPP; @@ -4474,39 +4451,24 @@ int ocfs2_reflink_ioctl(struct inode *inode, return error; } - error = ocfs2_user_path_parent(newname, &nd, &to); - if (error) { - mlog_errno(error); - goto out; - } - - error = -EXDEV; - if (old_path.mnt != nd.path.mnt) - goto out_release; - new_dentry = lookup_create(&nd, 0); + new_dentry = user_path_create(AT_FDCWD, newname, &new_path, 0); error = PTR_ERR(new_dentry); if (IS_ERR(new_dentry)) { mlog_errno(error); - goto out_unlock; + goto out; } - error = mnt_want_write(nd.path.mnt); - if (error) { + error = -EXDEV; + if (old_path.mnt != new_path.mnt) { mlog_errno(error); goto out_dput; } error = ocfs2_vfs_reflink(old_path.dentry, - nd.path.dentry->d_inode, + new_path.dentry->d_inode, new_dentry, preserve); - mnt_drop_write(nd.path.mnt); out_dput: - dput(new_dentry); -out_unlock: - mutex_unlock(&nd.path.dentry->d_inode->i_mutex); -out_release: - path_put(&nd.path); - putname(to); + done_path_create(&new_path, new_dentry); out: path_put(&old_path); diff --git a/fs/ocfs2/refcounttree.h b/fs/ocfs2/refcounttree.h index c8ce46f7d8e..6422bbcdb52 100644 --- a/fs/ocfs2/refcounttree.h +++ b/fs/ocfs2/refcounttree.h @@ -53,7 +53,7 @@ int ocfs2_prepare_refcount_change_for_del(struct inode *inode, int *credits, int *ref_blocks); int ocfs2_refcount_cow(struct inode *inode, - struct file *filep, struct buffer_head *di_bh, + struct buffer_head *di_bh, u32 cpos, u32 write_len, u32 max_cpos); typedef int (ocfs2_post_refcount_func)(struct inode *inode, @@ -84,6 +84,17 @@ int ocfs2_refcount_cow_xattr(struct inode *inode, struct buffer_head *ref_root_bh, u32 cpos, u32 write_len, struct ocfs2_post_refcount *post); +int ocfs2_duplicate_clusters_by_page(handle_t *handle, + struct inode *inode, + u32 cpos, u32 old_cluster, + u32 new_cluster, u32 new_len); +int ocfs2_duplicate_clusters_by_jbd(handle_t *handle, + struct inode *inode, + u32 cpos, u32 old_cluster, + u32 new_cluster, u32 new_len); +int ocfs2_cow_sync_writeback(struct super_block *sb, + struct inode *inode, + u32 cpos, u32 num_clusters); int ocfs2_add_refcount_flag(struct inode *inode, struct ocfs2_extent_tree *data_et, struct ocfs2_caching_info *ref_ci, diff --git a/fs/ocfs2/resize.c b/fs/ocfs2/resize.c index ec55add7604..d5da6f62414 100644 --- a/fs/ocfs2/resize.c +++ b/fs/ocfs2/resize.c @@ -53,8 +53,6 @@ */ static u16 ocfs2_calc_new_backup_super(struct inode *inode, struct ocfs2_group_desc *gd, - int new_clusters, - u32 first_new_cluster, u16 cl_cpg, int set) { @@ -127,8 +125,6 @@ static int ocfs2_update_last_group_and_inode(handle_t *handle, OCFS2_FEATURE_COMPAT_BACKUP_SB)) { backups = ocfs2_calc_new_backup_super(bm_inode, group, - new_clusters, - first_new_cluster, cl_cpg, 1); le16_add_cpu(&group->bg_free_bits_count, -1 * backups); } @@ -157,7 +153,7 @@ static int ocfs2_update_last_group_and_inode(handle_t *handle, spin_lock(&OCFS2_I(bm_inode)->ip_lock); OCFS2_I(bm_inode)->ip_clusters = le32_to_cpu(fe->i_clusters); - le64_add_cpu(&fe->i_size, new_clusters << osb->s_clustersize_bits); + le64_add_cpu(&fe->i_size, (u64)new_clusters << osb->s_clustersize_bits); spin_unlock(&OCFS2_I(bm_inode)->ip_lock); i_size_write(bm_inode, le64_to_cpu(fe->i_size)); @@ -167,8 +163,6 @@ out_rollback: if (ret < 0) { ocfs2_calc_new_backup_super(bm_inode, group, - new_clusters, - first_new_cluster, cl_cpg, 0); le16_add_cpu(&group->bg_free_bits_count, backups); le16_add_cpu(&group->bg_bits, -1 * num_bits); @@ -469,6 +463,7 @@ int ocfs2_group_add(struct inode *inode, struct ocfs2_new_group_input *input) struct ocfs2_chain_list *cl; struct ocfs2_chain_rec *cr; u16 cl_bpc; + u64 bg_ptr; if (ocfs2_is_hard_readonly(osb) || ocfs2_is_soft_readonly(osb)) return -EROFS; @@ -513,7 +508,7 @@ int ocfs2_group_add(struct inode *inode, struct ocfs2_new_group_input *input) ret = ocfs2_verify_group_and_input(main_bm_inode, fe, input, group_bh); if (ret) { mlog_errno(ret); - goto out_unlock; + goto out_free_group_bh; } trace_ocfs2_group_add((unsigned long long)input->group, @@ -523,7 +518,7 @@ int ocfs2_group_add(struct inode *inode, struct ocfs2_new_group_input *input) if (IS_ERR(handle)) { mlog_errno(PTR_ERR(handle)); ret = -EINVAL; - goto out_unlock; + goto out_free_group_bh; } cl_bpc = le16_to_cpu(fe->id2.i_chain.cl_bpc); @@ -538,12 +533,14 @@ int ocfs2_group_add(struct inode *inode, struct ocfs2_new_group_input *input) } group = (struct ocfs2_group_desc *)group_bh->b_data; + bg_ptr = le64_to_cpu(group->bg_next_group); group->bg_next_group = cr->c_blkno; ocfs2_journal_dirty(handle, group_bh); ret = ocfs2_journal_access_di(handle, INODE_CACHE(main_bm_inode), main_bm_bh, OCFS2_JOURNAL_ACCESS_WRITE); if (ret < 0) { + group->bg_next_group = cpu_to_le64(bg_ptr); mlog_errno(ret); goto out_commit; } @@ -566,7 +563,7 @@ int ocfs2_group_add(struct inode *inode, struct ocfs2_new_group_input *input) spin_lock(&OCFS2_I(main_bm_inode)->ip_lock); OCFS2_I(main_bm_inode)->ip_clusters = le32_to_cpu(fe->i_clusters); - le64_add_cpu(&fe->i_size, input->clusters << osb->s_clustersize_bits); + le64_add_cpu(&fe->i_size, (u64)input->clusters << osb->s_clustersize_bits); spin_unlock(&OCFS2_I(main_bm_inode)->ip_lock); i_size_write(main_bm_inode, le64_to_cpu(fe->i_size)); @@ -574,8 +571,11 @@ int ocfs2_group_add(struct inode *inode, struct ocfs2_new_group_input *input) out_commit: ocfs2_commit_trans(osb, handle); -out_unlock: + +out_free_group_bh: brelse(group_bh); + +out_unlock: brelse(main_bm_bh); ocfs2_inode_unlock(main_bm_inode, 1); diff --git a/fs/ocfs2/slot_map.c b/fs/ocfs2/slot_map.c index 26fc0014d50..1424c151ccc 100644 --- a/fs/ocfs2/slot_map.c +++ b/fs/ocfs2/slot_map.c @@ -493,8 +493,8 @@ int ocfs2_find_slot(struct ocfs2_super *osb) goto bail; } } else - mlog(ML_NOTICE, "slot %d is already allocated to this node!\n", - slot); + printk(KERN_INFO "ocfs2: Slot %d on device (%s) was already " + "allocated to this node!\n", slot, osb->dev_str); ocfs2_set_slot(si, slot, osb->node_num); osb->slot_num = slot; diff --git a/fs/ocfs2/stack_o2cb.c b/fs/ocfs2/stack_o2cb.c index 19965b00c43..1724d43d3da 100644 --- a/fs/ocfs2/stack_o2cb.c +++ b/fs/ocfs2/stack_o2cb.c @@ -28,6 +28,7 @@ #include "cluster/masklog.h" #include "cluster/nodemanager.h" #include "cluster/heartbeat.h" +#include "cluster/tcp.h" #include "stackglue.h" @@ -256,6 +257,61 @@ static void o2cb_dump_lksb(struct ocfs2_dlm_lksb *lksb) } /* + * Check if this node is heartbeating and is connected to all other + * heartbeating nodes. + */ +static int o2cb_cluster_check(void) +{ + u8 node_num; + int i; + unsigned long hbmap[BITS_TO_LONGS(O2NM_MAX_NODES)]; + unsigned long netmap[BITS_TO_LONGS(O2NM_MAX_NODES)]; + + node_num = o2nm_this_node(); + if (node_num == O2NM_MAX_NODES) { + printk(KERN_ERR "o2cb: This node has not been configured.\n"); + return -EINVAL; + } + + /* + * o2dlm expects o2net sockets to be created. If not, then + * dlm_join_domain() fails with a stack of errors which are both cryptic + * and incomplete. The idea here is to detect upfront whether we have + * managed to connect to all nodes or not. If not, then list the nodes + * to allow the user to check the configuration (incorrect IP, firewall, + * etc.) Yes, this is racy. But its not the end of the world. + */ +#define O2CB_MAP_STABILIZE_COUNT 60 + for (i = 0; i < O2CB_MAP_STABILIZE_COUNT; ++i) { + o2hb_fill_node_map(hbmap, sizeof(hbmap)); + if (!test_bit(node_num, hbmap)) { + printk(KERN_ERR "o2cb: %s heartbeat has not been " + "started.\n", (o2hb_global_heartbeat_active() ? + "Global" : "Local")); + return -EINVAL; + } + o2net_fill_node_map(netmap, sizeof(netmap)); + /* Force set the current node to allow easy compare */ + set_bit(node_num, netmap); + if (!memcmp(hbmap, netmap, sizeof(hbmap))) + return 0; + if (i < O2CB_MAP_STABILIZE_COUNT) + msleep(1000); + } + + printk(KERN_ERR "o2cb: This node could not connect to nodes:"); + i = -1; + while ((i = find_next_bit(hbmap, O2NM_MAX_NODES, + i + 1)) < O2NM_MAX_NODES) { + if (!test_bit(i, netmap)) + printk(" %u", i); + } + printk(".\n"); + + return -ENOTCONN; +} + +/* * Called from the dlm when it's about to evict a node. This is how the * classic stack signals node death. */ @@ -263,8 +319,8 @@ static void o2dlm_eviction_cb(int node_num, void *data) { struct ocfs2_cluster_connection *conn = data; - mlog(ML_NOTICE, "o2dlm has evicted node %d from group %.*s\n", - node_num, conn->cc_namelen, conn->cc_name); + printk(KERN_NOTICE "o2cb: o2dlm has evicted node %d from domain %.*s\n", + node_num, conn->cc_namelen, conn->cc_name); conn->cc_recovery_handler(node_num, conn->cc_recovery_data); } @@ -280,12 +336,11 @@ static int o2cb_cluster_connect(struct ocfs2_cluster_connection *conn) BUG_ON(conn == NULL); BUG_ON(conn->cc_proto == NULL); - /* for now we only have one cluster/node, make sure we see it - * in the heartbeat universe */ - if (!o2hb_check_local_node_heartbeating()) { - if (o2hb_global_heartbeat_active()) - mlog(ML_ERROR, "Global heartbeat not started\n"); - rc = -EINVAL; + /* Ensure cluster stack is up and all nodes are connected */ + rc = o2cb_cluster_check(); + if (rc) { + printk(KERN_ERR "o2cb: Cluster check failed. Fix errors " + "before retrying.\n"); goto out; } @@ -321,7 +376,7 @@ static int o2cb_cluster_connect(struct ocfs2_cluster_connection *conn) dlm_register_eviction_cb(dlm, &priv->op_eviction_cb); out_free: - if (rc && conn->cc_private) + if (rc) kfree(conn->cc_private); out: @@ -343,7 +398,8 @@ static int o2cb_cluster_disconnect(struct ocfs2_cluster_connection *conn) return 0; } -static int o2cb_cluster_this_node(unsigned int *node) +static int o2cb_cluster_this_node(struct ocfs2_cluster_connection *conn, + unsigned int *node) { int node_num; diff --git a/fs/ocfs2/stack_user.c b/fs/ocfs2/stack_user.c index a5ebe421195..13a8537d8e8 100644 --- a/fs/ocfs2/stack_user.c +++ b/fs/ocfs2/stack_user.c @@ -23,6 +23,7 @@ #include <linux/mutex.h> #include <linux/slab.h> #include <linux/reboot.h> +#include <linux/sched.h> #include <asm/uaccess.h> #include "stackglue.h" @@ -102,6 +103,12 @@ #define OCFS2_TEXT_UUID_LEN 32 #define OCFS2_CONTROL_MESSAGE_VERNUM_LEN 2 #define OCFS2_CONTROL_MESSAGE_NODENUM_LEN 8 +#define VERSION_LOCK "version_lock" + +enum ocfs2_connection_type { + WITH_CONTROLD, + NO_CONTROLD +}; /* * ocfs2_live_connection is refcounted because the filesystem and @@ -110,6 +117,13 @@ struct ocfs2_live_connection { struct list_head oc_list; struct ocfs2_cluster_connection *oc_conn; + enum ocfs2_connection_type oc_type; + atomic_t oc_this_node; + int oc_our_slot; + struct dlm_lksb oc_version_lksb; + char oc_lvb[DLM_LVB_LEN]; + struct completion oc_sync_wait; + wait_queue_head_t oc_wait; }; struct ocfs2_control_private { @@ -198,20 +212,15 @@ static struct ocfs2_live_connection *ocfs2_connection_find(const char *name) * mount path. Since the VFS prevents multiple calls to * fill_super(), we can't get dupes here. */ -static int ocfs2_live_connection_new(struct ocfs2_cluster_connection *conn, - struct ocfs2_live_connection **c_ret) +static int ocfs2_live_connection_attach(struct ocfs2_cluster_connection *conn, + struct ocfs2_live_connection *c) { int rc = 0; - struct ocfs2_live_connection *c; - - c = kzalloc(sizeof(struct ocfs2_live_connection), GFP_KERNEL); - if (!c) - return -ENOMEM; mutex_lock(&ocfs2_control_lock); c->oc_conn = conn; - if (atomic_read(&ocfs2_control_opened)) + if ((c->oc_type == NO_CONTROLD) || atomic_read(&ocfs2_control_opened)) list_add(&c->oc_list, &ocfs2_live_connection_list); else { printk(KERN_ERR @@ -220,12 +229,6 @@ static int ocfs2_live_connection_new(struct ocfs2_cluster_connection *conn, } mutex_unlock(&ocfs2_control_lock); - - if (!rc) - *c_ret = c; - else - kfree(c); - return rc; } @@ -799,18 +802,251 @@ static int fs_protocol_compare(struct ocfs2_protocol_version *existing, return 0; } +static void lvb_to_version(char *lvb, struct ocfs2_protocol_version *ver) +{ + struct ocfs2_protocol_version *pv = + (struct ocfs2_protocol_version *)lvb; + /* + * ocfs2_protocol_version has two u8 variables, so we don't + * need any endian conversion. + */ + ver->pv_major = pv->pv_major; + ver->pv_minor = pv->pv_minor; +} + +static void version_to_lvb(struct ocfs2_protocol_version *ver, char *lvb) +{ + struct ocfs2_protocol_version *pv = + (struct ocfs2_protocol_version *)lvb; + /* + * ocfs2_protocol_version has two u8 variables, so we don't + * need any endian conversion. + */ + pv->pv_major = ver->pv_major; + pv->pv_minor = ver->pv_minor; +} + +static void sync_wait_cb(void *arg) +{ + struct ocfs2_cluster_connection *conn = arg; + struct ocfs2_live_connection *lc = conn->cc_private; + complete(&lc->oc_sync_wait); +} + +static int sync_unlock(struct ocfs2_cluster_connection *conn, + struct dlm_lksb *lksb, char *name) +{ + int error; + struct ocfs2_live_connection *lc = conn->cc_private; + + error = dlm_unlock(conn->cc_lockspace, lksb->sb_lkid, 0, lksb, conn); + if (error) { + printk(KERN_ERR "%s lkid %x error %d\n", + name, lksb->sb_lkid, error); + return error; + } + + wait_for_completion(&lc->oc_sync_wait); + + if (lksb->sb_status != -DLM_EUNLOCK) { + printk(KERN_ERR "%s lkid %x status %d\n", + name, lksb->sb_lkid, lksb->sb_status); + return -1; + } + return 0; +} + +static int sync_lock(struct ocfs2_cluster_connection *conn, + int mode, uint32_t flags, + struct dlm_lksb *lksb, char *name) +{ + int error, status; + struct ocfs2_live_connection *lc = conn->cc_private; + + error = dlm_lock(conn->cc_lockspace, mode, lksb, flags, + name, strlen(name), + 0, sync_wait_cb, conn, NULL); + if (error) { + printk(KERN_ERR "%s lkid %x flags %x mode %d error %d\n", + name, lksb->sb_lkid, flags, mode, error); + return error; + } + + wait_for_completion(&lc->oc_sync_wait); + + status = lksb->sb_status; + + if (status && status != -EAGAIN) { + printk(KERN_ERR "%s lkid %x flags %x mode %d status %d\n", + name, lksb->sb_lkid, flags, mode, status); + } + + return status; +} + + +static int version_lock(struct ocfs2_cluster_connection *conn, int mode, + int flags) +{ + struct ocfs2_live_connection *lc = conn->cc_private; + return sync_lock(conn, mode, flags, + &lc->oc_version_lksb, VERSION_LOCK); +} + +static int version_unlock(struct ocfs2_cluster_connection *conn) +{ + struct ocfs2_live_connection *lc = conn->cc_private; + return sync_unlock(conn, &lc->oc_version_lksb, VERSION_LOCK); +} + +/* get_protocol_version() + * + * To exchange ocfs2 versioning, we use the LVB of the version dlm lock. + * The algorithm is: + * 1. Attempt to take the lock in EX mode (non-blocking). + * 2. If successful (which means it is the first mount), write the + * version number and downconvert to PR lock. + * 3. If unsuccessful (returns -EAGAIN), read the version from the LVB after + * taking the PR lock. + */ + +static int get_protocol_version(struct ocfs2_cluster_connection *conn) +{ + int ret; + struct ocfs2_live_connection *lc = conn->cc_private; + struct ocfs2_protocol_version pv; + + running_proto.pv_major = + ocfs2_user_plugin.sp_max_proto.pv_major; + running_proto.pv_minor = + ocfs2_user_plugin.sp_max_proto.pv_minor; + + lc->oc_version_lksb.sb_lvbptr = lc->oc_lvb; + ret = version_lock(conn, DLM_LOCK_EX, + DLM_LKF_VALBLK|DLM_LKF_NOQUEUE); + if (!ret) { + conn->cc_version.pv_major = running_proto.pv_major; + conn->cc_version.pv_minor = running_proto.pv_minor; + version_to_lvb(&running_proto, lc->oc_lvb); + version_lock(conn, DLM_LOCK_PR, DLM_LKF_CONVERT|DLM_LKF_VALBLK); + } else if (ret == -EAGAIN) { + ret = version_lock(conn, DLM_LOCK_PR, DLM_LKF_VALBLK); + if (ret) + goto out; + lvb_to_version(lc->oc_lvb, &pv); + + if ((pv.pv_major != running_proto.pv_major) || + (pv.pv_minor > running_proto.pv_minor)) { + ret = -EINVAL; + goto out; + } + + conn->cc_version.pv_major = pv.pv_major; + conn->cc_version.pv_minor = pv.pv_minor; + } +out: + return ret; +} + +static void user_recover_prep(void *arg) +{ +} + +static void user_recover_slot(void *arg, struct dlm_slot *slot) +{ + struct ocfs2_cluster_connection *conn = arg; + printk(KERN_INFO "ocfs2: Node %d/%d down. Initiating recovery.\n", + slot->nodeid, slot->slot); + conn->cc_recovery_handler(slot->nodeid, conn->cc_recovery_data); + +} + +static void user_recover_done(void *arg, struct dlm_slot *slots, + int num_slots, int our_slot, + uint32_t generation) +{ + struct ocfs2_cluster_connection *conn = arg; + struct ocfs2_live_connection *lc = conn->cc_private; + int i; + + for (i = 0; i < num_slots; i++) + if (slots[i].slot == our_slot) { + atomic_set(&lc->oc_this_node, slots[i].nodeid); + break; + } + + lc->oc_our_slot = our_slot; + wake_up(&lc->oc_wait); +} + +static const struct dlm_lockspace_ops ocfs2_ls_ops = { + .recover_prep = user_recover_prep, + .recover_slot = user_recover_slot, + .recover_done = user_recover_done, +}; + +static int user_cluster_disconnect(struct ocfs2_cluster_connection *conn) +{ + version_unlock(conn); + dlm_release_lockspace(conn->cc_lockspace, 2); + conn->cc_lockspace = NULL; + ocfs2_live_connection_drop(conn->cc_private); + conn->cc_private = NULL; + return 0; +} + static int user_cluster_connect(struct ocfs2_cluster_connection *conn) { dlm_lockspace_t *fsdlm; - struct ocfs2_live_connection *uninitialized_var(control); - int rc = 0; + struct ocfs2_live_connection *lc; + int rc, ops_rv; BUG_ON(conn == NULL); - rc = ocfs2_live_connection_new(conn, &control); + lc = kzalloc(sizeof(struct ocfs2_live_connection), GFP_KERNEL); + if (!lc) { + rc = -ENOMEM; + goto out; + } + + init_waitqueue_head(&lc->oc_wait); + init_completion(&lc->oc_sync_wait); + atomic_set(&lc->oc_this_node, 0); + conn->cc_private = lc; + lc->oc_type = NO_CONTROLD; + + rc = dlm_new_lockspace(conn->cc_name, conn->cc_cluster_name, + DLM_LSFL_FS, DLM_LVB_LEN, + &ocfs2_ls_ops, conn, &ops_rv, &fsdlm); + if (rc) + goto out; + + if (ops_rv == -EOPNOTSUPP) { + lc->oc_type = WITH_CONTROLD; + printk(KERN_NOTICE "ocfs2: You seem to be using an older " + "version of dlm_controld and/or ocfs2-tools." + " Please consider upgrading.\n"); + } else if (ops_rv) { + rc = ops_rv; + goto out; + } + conn->cc_lockspace = fsdlm; + + rc = ocfs2_live_connection_attach(conn, lc); if (rc) goto out; + if (lc->oc_type == NO_CONTROLD) { + rc = get_protocol_version(conn); + if (rc) { + printk(KERN_ERR "ocfs2: Could not determine" + " locking version\n"); + user_cluster_disconnect(conn); + goto out; + } + wait_event(lc->oc_wait, (atomic_read(&lc->oc_this_node) > 0)); + } + /* * running_proto must have been set before we allowed any mounts * to proceed. @@ -818,42 +1054,34 @@ static int user_cluster_connect(struct ocfs2_cluster_connection *conn) if (fs_protocol_compare(&running_proto, &conn->cc_version)) { printk(KERN_ERR "Unable to mount with fs locking protocol version " - "%u.%u because the userspace control daemon has " - "negotiated %u.%u\n", + "%u.%u because negotiated protocol is %u.%u\n", conn->cc_version.pv_major, conn->cc_version.pv_minor, running_proto.pv_major, running_proto.pv_minor); rc = -EPROTO; - ocfs2_live_connection_drop(control); - goto out; - } - - rc = dlm_new_lockspace(conn->cc_name, strlen(conn->cc_name), - &fsdlm, DLM_LSFL_FS, DLM_LVB_LEN); - if (rc) { - ocfs2_live_connection_drop(control); - goto out; + ocfs2_live_connection_drop(lc); + lc = NULL; } - conn->cc_private = control; - conn->cc_lockspace = fsdlm; out: + if (rc && lc) + kfree(lc); return rc; } -static int user_cluster_disconnect(struct ocfs2_cluster_connection *conn) -{ - dlm_release_lockspace(conn->cc_lockspace, 2); - conn->cc_lockspace = NULL; - ocfs2_live_connection_drop(conn->cc_private); - conn->cc_private = NULL; - return 0; -} -static int user_cluster_this_node(unsigned int *this_node) +static int user_cluster_this_node(struct ocfs2_cluster_connection *conn, + unsigned int *this_node) { int rc; + struct ocfs2_live_connection *lc = conn->cc_private; + + if (lc->oc_type == WITH_CONTROLD) + rc = ocfs2_control_get_this_node(); + else if (lc->oc_type == NO_CONTROLD) + rc = atomic_read(&lc->oc_this_node); + else + rc = -EINVAL; - rc = ocfs2_control_get_this_node(); if (rc < 0) return rc; diff --git a/fs/ocfs2/stackglue.c b/fs/ocfs2/stackglue.c index 39abf89697e..5d965e83bd4 100644 --- a/fs/ocfs2/stackglue.c +++ b/fs/ocfs2/stackglue.c @@ -309,6 +309,8 @@ int ocfs2_plock(struct ocfs2_cluster_connection *conn, u64 ino, EXPORT_SYMBOL_GPL(ocfs2_plock); int ocfs2_cluster_connect(const char *stack_name, + const char *cluster_name, + int cluster_name_len, const char *group, int grouplen, struct ocfs2_locking_protocol *lproto, @@ -342,8 +344,12 @@ int ocfs2_cluster_connect(const char *stack_name, goto out; } - memcpy(new_conn->cc_name, group, grouplen); + strlcpy(new_conn->cc_name, group, GROUP_NAME_MAX + 1); new_conn->cc_namelen = grouplen; + if (cluster_name_len) + strlcpy(new_conn->cc_cluster_name, cluster_name, + CLUSTER_NAME_MAX + 1); + new_conn->cc_cluster_name_len = cluster_name_len; new_conn->cc_recovery_handler = recovery_handler; new_conn->cc_recovery_data = recovery_data; @@ -386,8 +392,9 @@ int ocfs2_cluster_connect_agnostic(const char *group, if (cluster_stack_name[0]) stack_name = cluster_stack_name; - return ocfs2_cluster_connect(stack_name, group, grouplen, lproto, - recovery_handler, recovery_data, conn); + return ocfs2_cluster_connect(stack_name, NULL, 0, group, grouplen, + lproto, recovery_handler, recovery_data, + conn); } EXPORT_SYMBOL_GPL(ocfs2_cluster_connect_agnostic); @@ -460,9 +467,10 @@ void ocfs2_cluster_hangup(const char *group, int grouplen) } EXPORT_SYMBOL_GPL(ocfs2_cluster_hangup); -int ocfs2_cluster_this_node(unsigned int *node) +int ocfs2_cluster_this_node(struct ocfs2_cluster_connection *conn, + unsigned int *node) { - return active_stack->sp_ops->this_node(node); + return active_stack->sp_ops->this_node(conn, node); } EXPORT_SYMBOL_GPL(ocfs2_cluster_this_node); @@ -488,7 +496,7 @@ static ssize_t ocfs2_max_locking_protocol_show(struct kobject *kobj, } static struct kobj_attribute ocfs2_attr_max_locking_protocol = - __ATTR(max_locking_protocol, S_IFREG | S_IRUGO, + __ATTR(max_locking_protocol, S_IRUGO, ocfs2_max_locking_protocol_show, NULL); static ssize_t ocfs2_loaded_cluster_plugins_show(struct kobject *kobj, @@ -520,7 +528,7 @@ static ssize_t ocfs2_loaded_cluster_plugins_show(struct kobject *kobj, } static struct kobj_attribute ocfs2_attr_loaded_cluster_plugins = - __ATTR(loaded_cluster_plugins, S_IFREG | S_IRUGO, + __ATTR(loaded_cluster_plugins, S_IRUGO, ocfs2_loaded_cluster_plugins_show, NULL); static ssize_t ocfs2_active_cluster_plugin_show(struct kobject *kobj, @@ -542,7 +550,7 @@ static ssize_t ocfs2_active_cluster_plugin_show(struct kobject *kobj, } static struct kobj_attribute ocfs2_attr_active_cluster_plugin = - __ATTR(active_cluster_plugin, S_IFREG | S_IRUGO, + __ATTR(active_cluster_plugin, S_IRUGO, ocfs2_active_cluster_plugin_show, NULL); static ssize_t ocfs2_cluster_stack_show(struct kobject *kobj, @@ -591,15 +599,29 @@ static ssize_t ocfs2_cluster_stack_store(struct kobject *kobj, static struct kobj_attribute ocfs2_attr_cluster_stack = - __ATTR(cluster_stack, S_IFREG | S_IRUGO | S_IWUSR, + __ATTR(cluster_stack, S_IRUGO | S_IWUSR, ocfs2_cluster_stack_show, ocfs2_cluster_stack_store); + + +static ssize_t ocfs2_dlm_recover_show(struct kobject *kobj, + struct kobj_attribute *attr, + char *buf) +{ + return snprintf(buf, PAGE_SIZE, "1\n"); +} + +static struct kobj_attribute ocfs2_attr_dlm_recover_support = + __ATTR(dlm_recover_callback_support, S_IRUGO, + ocfs2_dlm_recover_show, NULL); + static struct attribute *ocfs2_attrs[] = { &ocfs2_attr_max_locking_protocol.attr, &ocfs2_attr_loaded_cluster_plugins.attr, &ocfs2_attr_active_cluster_plugin.attr, &ocfs2_attr_cluster_stack.attr, + &ocfs2_attr_dlm_recover_support.attr, NULL, }; @@ -643,7 +665,7 @@ error: #define FS_OCFS2_NM 1 -static ctl_table ocfs2_nm_table[] = { +static struct ctl_table ocfs2_nm_table[] = { { .procname = "hb_ctl_path", .data = ocfs2_hb_ctl_path, @@ -654,7 +676,7 @@ static ctl_table ocfs2_nm_table[] = { { } }; -static ctl_table ocfs2_mod_table[] = { +static struct ctl_table ocfs2_mod_table[] = { { .procname = "nm", .data = NULL, @@ -665,7 +687,7 @@ static ctl_table ocfs2_mod_table[] = { { } }; -static ctl_table ocfs2_kern_table[] = { +static struct ctl_table ocfs2_kern_table[] = { { .procname = "ocfs2", .data = NULL, @@ -676,7 +698,7 @@ static ctl_table ocfs2_kern_table[] = { { } }; -static ctl_table ocfs2_root_table[] = { +static struct ctl_table ocfs2_root_table[] = { { .procname = "fs", .data = NULL, @@ -687,7 +709,7 @@ static ctl_table ocfs2_root_table[] = { { } }; -static struct ctl_table_header *ocfs2_table_header = NULL; +static struct ctl_table_header *ocfs2_table_header; /* diff --git a/fs/ocfs2/stackglue.h b/fs/ocfs2/stackglue.h index 1ec56fdb8d0..66334a30cea 100644 --- a/fs/ocfs2/stackglue.h +++ b/fs/ocfs2/stackglue.h @@ -45,6 +45,9 @@ struct file_lock; */ #define GROUP_NAME_MAX 64 +/* This shadows OCFS2_CLUSTER_NAME_LEN */ +#define CLUSTER_NAME_MAX 16 + /* * ocfs2_protocol_version changes when ocfs2 does something different in @@ -97,8 +100,10 @@ struct ocfs2_locking_protocol { * locking compatibility. */ struct ocfs2_cluster_connection { - char cc_name[GROUP_NAME_MAX]; + char cc_name[GROUP_NAME_MAX + 1]; int cc_namelen; + char cc_cluster_name[CLUSTER_NAME_MAX + 1]; + int cc_cluster_name_len; struct ocfs2_protocol_version cc_version; struct ocfs2_locking_protocol *cc_proto; void (*cc_recovery_handler)(int node_num, void *recovery_data); @@ -152,7 +157,8 @@ struct ocfs2_stack_operations { * ->this_node() returns the cluster's unique identifier for the * local node. */ - int (*this_node)(unsigned int *node); + int (*this_node)(struct ocfs2_cluster_connection *conn, + unsigned int *node); /* * Call the underlying dlm lock function. The ->dlm_lock() @@ -239,6 +245,8 @@ struct ocfs2_stack_plugin { /* Used by the filesystem */ int ocfs2_cluster_connect(const char *stack_name, + const char *cluster_name, + int cluster_name_len, const char *group, int grouplen, struct ocfs2_locking_protocol *lproto, @@ -260,7 +268,8 @@ int ocfs2_cluster_connect_agnostic(const char *group, int ocfs2_cluster_disconnect(struct ocfs2_cluster_connection *conn, int hangup_pending); void ocfs2_cluster_hangup(const char *group, int grouplen); -int ocfs2_cluster_this_node(unsigned int *node); +int ocfs2_cluster_this_node(struct ocfs2_cluster_connection *conn, + unsigned int *node); struct ocfs2_lock_res; int ocfs2_dlm_lock(struct ocfs2_cluster_connection *conn, diff --git a/fs/ocfs2/suballoc.c b/fs/ocfs2/suballoc.c index ba5d97e4a73..0cb889a17ae 100644 --- a/fs/ocfs2/suballoc.c +++ b/fs/ocfs2/suballoc.c @@ -113,12 +113,6 @@ static int ocfs2_claim_suballoc_bits(struct ocfs2_alloc_context *ac, struct ocfs2_suballoc_result *res); static int ocfs2_test_bg_bit_allocatable(struct buffer_head *bg_bh, int nr); -static inline int ocfs2_block_group_set_bits(handle_t *handle, - struct inode *alloc_inode, - struct ocfs2_group_desc *bg, - struct buffer_head *group_bh, - unsigned int bit_off, - unsigned int num_bits); static int ocfs2_relink_block_group(handle_t *handle, struct inode *alloc_inode, struct buffer_head *fe_bh, @@ -481,7 +475,7 @@ ocfs2_block_group_alloc_contig(struct ocfs2_super *osb, handle_t *handle, bg_bh = sb_getblk(osb->sb, bg_blkno); if (!bg_bh) { - status = -EIO; + status = -ENOMEM; mlog_errno(status); goto bail; } @@ -600,7 +594,7 @@ static void ocfs2_bg_alloc_cleanup(handle_t *handle, ret = ocfs2_free_clusters(handle, cluster_ac->ac_inode, cluster_ac->ac_bh, le64_to_cpu(rec->e_blkno), - le32_to_cpu(rec->e_leaf_clusters)); + le16_to_cpu(rec->e_leaf_clusters)); if (ret) mlog_errno(ret); /* Try all the clusters to free */ @@ -642,7 +636,7 @@ ocfs2_block_group_alloc_discontig(handle_t *handle, * cluster groups will be staying in cache for the duration of * this operation. */ - ac->ac_allow_chain_relink = 0; + ac->ac_disable_chain_relink = 1; /* Claim the first region */ status = ocfs2_block_group_claim_bits(osb, handle, ac, min_bits, @@ -661,7 +655,7 @@ ocfs2_block_group_alloc_discontig(handle_t *handle, bg_bh = sb_getblk(osb->sb, bg_blkno); if (!bg_bh) { - status = -EIO; + status = -ENOMEM; mlog_errno(status); goto bail; } @@ -777,6 +771,7 @@ static int ocfs2_block_group_alloc(struct ocfs2_super *osb, spin_unlock(&OCFS2_I(alloc_inode)->ip_lock); i_size_write(alloc_inode, le64_to_cpu(fe->i_size)); alloc_inode->i_blocks = ocfs2_inode_sector_count(alloc_inode); + ocfs2_update_inode_fsync_trans(handle, alloc_inode, 0); status = 0; @@ -1343,7 +1338,7 @@ static int ocfs2_block_group_find_clear_bits(struct ocfs2_super *osb, return status; } -static inline int ocfs2_block_group_set_bits(handle_t *handle, +int ocfs2_block_group_set_bits(handle_t *handle, struct inode *alloc_inode, struct ocfs2_group_desc *bg, struct buffer_head *group_bh, @@ -1388,8 +1383,6 @@ static inline int ocfs2_block_group_set_bits(handle_t *handle, ocfs2_journal_dirty(handle, group_bh); bail: - if (status) - mlog_errno(status); return status; } @@ -1422,7 +1415,7 @@ static int ocfs2_relink_block_group(handle_t *handle, int status; /* there is a really tiny chance the journal calls could fail, * but we wouldn't want inconsistent blocks in *any* case. */ - u64 fe_ptr, bg_ptr, prev_bg_ptr; + u64 bg_ptr, prev_bg_ptr; struct ocfs2_dinode *fe = (struct ocfs2_dinode *) fe_bh->b_data; struct ocfs2_group_desc *bg = (struct ocfs2_group_desc *) bg_bh->b_data; struct ocfs2_group_desc *prev_bg = (struct ocfs2_group_desc *) prev_bg_bh->b_data; @@ -1437,51 +1430,44 @@ static int ocfs2_relink_block_group(handle_t *handle, (unsigned long long)le64_to_cpu(bg->bg_blkno), (unsigned long long)le64_to_cpu(prev_bg->bg_blkno)); - fe_ptr = le64_to_cpu(fe->id2.i_chain.cl_recs[chain].c_blkno); bg_ptr = le64_to_cpu(bg->bg_next_group); prev_bg_ptr = le64_to_cpu(prev_bg->bg_next_group); status = ocfs2_journal_access_gd(handle, INODE_CACHE(alloc_inode), prev_bg_bh, OCFS2_JOURNAL_ACCESS_WRITE); - if (status < 0) { - mlog_errno(status); - goto out_rollback; - } + if (status < 0) + goto out; prev_bg->bg_next_group = bg->bg_next_group; ocfs2_journal_dirty(handle, prev_bg_bh); status = ocfs2_journal_access_gd(handle, INODE_CACHE(alloc_inode), bg_bh, OCFS2_JOURNAL_ACCESS_WRITE); - if (status < 0) { - mlog_errno(status); - goto out_rollback; - } + if (status < 0) + goto out_rollback_prev_bg; bg->bg_next_group = fe->id2.i_chain.cl_recs[chain].c_blkno; ocfs2_journal_dirty(handle, bg_bh); status = ocfs2_journal_access_di(handle, INODE_CACHE(alloc_inode), fe_bh, OCFS2_JOURNAL_ACCESS_WRITE); - if (status < 0) { - mlog_errno(status); - goto out_rollback; - } + if (status < 0) + goto out_rollback_bg; fe->id2.i_chain.cl_recs[chain].c_blkno = bg->bg_blkno; ocfs2_journal_dirty(handle, fe_bh); -out_rollback: - if (status < 0) { - fe->id2.i_chain.cl_recs[chain].c_blkno = cpu_to_le64(fe_ptr); - bg->bg_next_group = cpu_to_le64(bg_ptr); - prev_bg->bg_next_group = cpu_to_le64(prev_bg_ptr); - } - - if (status) +out: + if (status < 0) mlog_errno(status); return status; + +out_rollback_bg: + bg->bg_next_group = cpu_to_le64(bg_ptr); +out_rollback_prev_bg: + prev_bg->bg_next_group = cpu_to_le64(prev_bg_ptr); + goto out; } static inline int ocfs2_block_group_reasonably_empty(struct ocfs2_group_desc *bg, @@ -1595,7 +1581,7 @@ static int ocfs2_block_group_search(struct inode *inode, return ret; } -static int ocfs2_alloc_dinode_update_counts(struct inode *inode, +int ocfs2_alloc_dinode_update_counts(struct inode *inode, handle_t *handle, struct buffer_head *di_bh, u32 num_bits, @@ -1622,13 +1608,28 @@ out: return ret; } +void ocfs2_rollback_alloc_dinode_counts(struct inode *inode, + struct buffer_head *di_bh, + u32 num_bits, + u16 chain) +{ + u32 tmp_used; + struct ocfs2_dinode *di = (struct ocfs2_dinode *) di_bh->b_data; + struct ocfs2_chain_list *cl; + + cl = (struct ocfs2_chain_list *)&di->id2.i_chain; + tmp_used = le32_to_cpu(di->id1.bitmap1.i_used); + di->id1.bitmap1.i_used = cpu_to_le32(tmp_used - num_bits); + le32_add_cpu(&cl->cl_recs[chain].c_free, num_bits); +} + static int ocfs2_bg_discontig_fix_by_rec(struct ocfs2_suballoc_result *res, struct ocfs2_extent_rec *rec, struct ocfs2_chain_list *cl) { unsigned int bpc = le16_to_cpu(cl->cl_bpc); unsigned int bitoff = le32_to_cpu(rec->e_cpos) * bpc; - unsigned int bitcount = le32_to_cpu(rec->e_leaf_clusters) * bpc; + unsigned int bitcount = le16_to_cpu(rec->e_leaf_clusters) * bpc; if (res->sr_bit_offset < bitoff) return 0; @@ -1722,8 +1723,12 @@ static int ocfs2_search_one_group(struct ocfs2_alloc_context *ac, ret = ocfs2_block_group_set_bits(handle, alloc_inode, gd, group_bh, res->sr_bit_offset, res->sr_bits); - if (ret < 0) + if (ret < 0) { + ocfs2_rollback_alloc_dinode_counts(alloc_inode, ac->ac_bh, + res->sr_bits, + le16_to_cpu(gd->bg_chain)); mlog_errno(ret); + } out_loc_only: *bits_left = le16_to_cpu(gd->bg_free_bits_count); @@ -1823,7 +1828,7 @@ static int ocfs2_search_chain(struct ocfs2_alloc_context *ac, * Do this *after* figuring out how many bits we're taking out * of our target group. */ - if (ac->ac_allow_chain_relink && + if (!ac->ac_disable_chain_relink && (prev_group_bh) && (ocfs2_block_group_reasonably_empty(bg, res->sr_bits))) { status = ocfs2_relink_block_group(handle, alloc_inode, @@ -1853,6 +1858,8 @@ static int ocfs2_search_chain(struct ocfs2_alloc_context *ac, res->sr_bit_offset, res->sr_bits); if (status < 0) { + ocfs2_rollback_alloc_dinode_counts(alloc_inode, + ac->ac_bh, res->sr_bits, chain); mlog_errno(status); goto bail; } @@ -1928,7 +1935,6 @@ static int ocfs2_claim_suballoc_bits(struct ocfs2_alloc_context *ac, victim = ocfs2_find_victim_chain(cl); ac->ac_chain = victim; - ac->ac_allow_chain_relink = 1; status = ocfs2_search_chain(ac, handle, bits_wanted, min_bits, res, &bits_left); @@ -1947,7 +1953,7 @@ static int ocfs2_claim_suballoc_bits(struct ocfs2_alloc_context *ac, * searching each chain in order. Don't allow chain relinking * because we only calculate enough journal credits for one * relink per alloc. */ - ac->ac_allow_chain_relink = 0; + ac->ac_disable_chain_relink = 1; for (i = 0; i < le16_to_cpu(cl->cl_next_free_rec); i ++) { if (i == victim) continue; @@ -2107,7 +2113,7 @@ int ocfs2_find_new_inode_loc(struct inode *dir, ac->ac_find_loc_priv = res; *fe_blkno = res->sr_blkno; - + ocfs2_update_inode_fsync_trans(handle, dir, 0); out: if (handle) ocfs2_commit_trans(OCFS2_SB(dir->i_sb), handle); @@ -2165,6 +2171,8 @@ int ocfs2_claim_new_inode_at_loc(handle_t *handle, res->sr_bit_offset, res->sr_bits); if (ret < 0) { + ocfs2_rollback_alloc_dinode_counts(ac->ac_inode, + ac->ac_bh, res->sr_bits, chain); mlog_errno(ret); goto out; } @@ -2886,6 +2894,7 @@ int ocfs2_test_inode_bit(struct ocfs2_super *osb, u64 blkno, int *res) status = ocfs2_inode_lock(inode_alloc_inode, &alloc_bh, 0); if (status < 0) { mutex_unlock(&inode_alloc_inode->i_mutex); + iput(inode_alloc_inode); mlog(ML_ERROR, "lock on alloc inode on slot %u failed %d\n", (u32)suballoc_slot, status); goto bail; diff --git a/fs/ocfs2/suballoc.h b/fs/ocfs2/suballoc.h index b8afabfeede..2d2501767c0 100644 --- a/fs/ocfs2/suballoc.h +++ b/fs/ocfs2/suballoc.h @@ -49,7 +49,7 @@ struct ocfs2_alloc_context { /* these are used by the chain search */ u16 ac_chain; - int ac_allow_chain_relink; + int ac_disable_chain_relink; group_search_t *ac_group_search; u64 ac_last_group; @@ -86,6 +86,22 @@ int ocfs2_reserve_clusters(struct ocfs2_super *osb, u32 bits_wanted, struct ocfs2_alloc_context **ac); +int ocfs2_alloc_dinode_update_counts(struct inode *inode, + handle_t *handle, + struct buffer_head *di_bh, + u32 num_bits, + u16 chain); +void ocfs2_rollback_alloc_dinode_counts(struct inode *inode, + struct buffer_head *di_bh, + u32 num_bits, + u16 chain); +int ocfs2_block_group_set_bits(handle_t *handle, + struct inode *alloc_inode, + struct ocfs2_group_desc *bg, + struct buffer_head *group_bh, + unsigned int bit_off, + unsigned int num_bits); + int ocfs2_claim_metadata(handle_t *handle, struct ocfs2_alloc_context *ac, u32 bits_wanted, diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c index 5a521c74885..ddb662b3244 100644 --- a/fs/ocfs2/super.c +++ b/fs/ocfs2/super.c @@ -41,6 +41,7 @@ #include <linux/mount.h> #include <linux/seq_file.h> #include <linux/quotaops.h> +#include <linux/cleancache.h> #define CREATE_TRACE_POINTS #include "ocfs2_trace.h" @@ -53,6 +54,7 @@ #include "ocfs1_fs_compat.h" #include "alloc.h" +#include "aops.h" #include "blockcheck.h" #include "dlmglue.h" #include "export.h" @@ -66,7 +68,6 @@ #include "super.h" #include "sysfile.h" #include "uptodate.h" -#include "ver.h" #include "xattr.h" #include "quota.h" #include "refcounttree.h" @@ -74,7 +75,7 @@ #include "buffer_head_io.h" -static struct kmem_cache *ocfs2_inode_cachep = NULL; +static struct kmem_cache *ocfs2_inode_cachep; struct kmem_cache *ocfs2_dquot_cachep; struct kmem_cache *ocfs2_qf_chunk_cachep; @@ -84,10 +85,11 @@ struct kmem_cache *ocfs2_qf_chunk_cachep; * workqueue and schedule on our own. */ struct workqueue_struct *ocfs2_wq = NULL; -static struct dentry *ocfs2_debugfs_root = NULL; +static struct dentry *ocfs2_debugfs_root; MODULE_AUTHOR("Oracle"); MODULE_LICENSE("GPL"); +MODULE_DESCRIPTION("OCFS2 cluster file system"); struct mount_options { @@ -106,7 +108,7 @@ static int ocfs2_parse_options(struct super_block *sb, char *options, int is_remount); static int ocfs2_check_set_options(struct super_block *sb, struct mount_options *options); -static int ocfs2_show_options(struct seq_file *s, struct vfsmount *mnt); +static int ocfs2_show_options(struct seq_file *s, struct dentry *root); static void ocfs2_put_super(struct super_block *sb); static int ocfs2_mount_volume(struct super_block *sb); static int ocfs2_remount(struct super_block *sb, int *flags, char *data); @@ -284,10 +286,9 @@ static int ocfs2_osb_dump(struct ocfs2_super *osb, char *buf, int len) spin_unlock(&osb->osb_lock); out += snprintf(buf + out, len - out, - "%10s => Pid: %d Interval: %lu Needs: %d\n", "Commit", + "%10s => Pid: %d Interval: %lu\n", "Commit", (osb->commit_task ? task_pid_nr(osb->commit_task) : -1), - osb->osb_commit_interval, - atomic_read(&osb->needs_checkpoint)); + osb->osb_commit_interval); out += snprintf(buf + out, len - out, "%10s => State: %d TxnId: %lu NumTxns: %d\n", @@ -560,6 +561,9 @@ static struct inode *ocfs2_alloc_inode(struct super_block *sb) if (!oi) return NULL; + oi->i_sync_tid = 0; + oi->i_datasync_tid = 0; + jbd2_journal_init_jbd_inode(&oi->ip_jinode, &oi->vfs_inode); return &oi->vfs_inode; } @@ -567,7 +571,6 @@ static struct inode *ocfs2_alloc_inode(struct super_block *sb) static void ocfs2_i_callback(struct rcu_head *head) { struct inode *inode = container_of(head, struct inode, i_rcu); - INIT_LIST_HEAD(&inode->i_dentry); kmem_cache_free(ocfs2_inode_cachep, OCFS2_I(inode)); } @@ -631,6 +634,8 @@ static int ocfs2_remount(struct super_block *sb, int *flags, char *data) struct ocfs2_super *osb = OCFS2_SB(sb); u32 tmp; + sync_filesystem(sb); + if (!ocfs2_parse_options(sb, data, &parsed_options, 1) || !ocfs2_check_set_options(sb, &parsed_options)) { ret = -EINVAL; @@ -1022,7 +1027,7 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent) struct inode *inode = NULL; struct ocfs2_super *osb = NULL; struct buffer_head *bh = NULL; - char nodestr[8]; + char nodestr[12]; struct ocfs2_blockcheck_stats stats; trace_ocfs2_fill_super(sb, data, silent); @@ -1071,7 +1076,7 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent) sb->s_magic = OCFS2_SUPER_MAGIC; - sb->s_flags = (sb->s_flags & ~MS_POSIXACL) | + sb->s_flags = (sb->s_flags & ~(MS_POSIXACL | MS_NOSEC)) | ((osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL) ? MS_POSIXACL : 0); /* Hard readonly mode only if: bdev_read_only, MS_RDONLY, @@ -1106,9 +1111,9 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent) ocfs2_set_ro_flag(osb, 1); - printk(KERN_NOTICE "Readonly device detected. No cluster " - "services will be utilized for this mount. Recovery " - "will be skipped.\n"); + printk(KERN_NOTICE "ocfs2: Readonly device (%s) detected. " + "Cluster services will not be used for this mount. " + "Recovery will be skipped.\n", osb->dev_str); } if (!ocfs2_is_hard_readonly(osb)) { @@ -1153,19 +1158,19 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent) } status = ocfs2_mount_volume(sb); - if (osb->root_inode) - inode = igrab(osb->root_inode); - if (status < 0) goto read_super_error; + if (osb->root_inode) + inode = igrab(osb->root_inode); + if (!inode) { status = -EIO; mlog_errno(status); goto read_super_error; } - root = d_alloc_root(inode); + root = d_make_root(inode); if (!root) { status = -ENOMEM; mlog_errno(status); @@ -1219,9 +1224,6 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent) read_super_error: brelse(bh); - if (inode) - iput(inode); - if (osb) { atomic_set(&osb->vol_state, VOLUME_DISABLED); wake_up(&osb->osb_mount_event); @@ -1241,33 +1243,15 @@ static struct dentry *ocfs2_mount(struct file_system_type *fs_type, return mount_bdev(fs_type, flags, dev_name, data, ocfs2_fill_super); } -static void ocfs2_kill_sb(struct super_block *sb) -{ - struct ocfs2_super *osb = OCFS2_SB(sb); - - /* Failed mount? */ - if (!osb || atomic_read(&osb->vol_state) == VOLUME_DISABLED) - goto out; - - /* Prevent further queueing of inode drop events */ - spin_lock(&dentry_list_lock); - ocfs2_set_osb_flag(osb, OCFS2_OSB_DROP_DENTRY_LOCK_IMMED); - spin_unlock(&dentry_list_lock); - /* Wait for work to finish and/or remove it */ - cancel_work_sync(&osb->dentry_lock_work); -out: - kill_block_super(sb); -} - static struct file_system_type ocfs2_fs_type = { .owner = THIS_MODULE, .name = "ocfs2", .mount = ocfs2_mount, - .kill_sb = ocfs2_kill_sb, - + .kill_sb = kill_block_super, .fs_flags = FS_REQUIRES_DEV|FS_RENAME_DOES_D_MOVE, .next = NULL }; +MODULE_ALIAS_FS("ocfs2"); static int ocfs2_check_set_options(struct super_block *sb, struct mount_options *options) @@ -1532,9 +1516,9 @@ bail: return status; } -static int ocfs2_show_options(struct seq_file *s, struct vfsmount *mnt) +static int ocfs2_show_options(struct seq_file *s, struct dentry *root) { - struct ocfs2_super *osb = OCFS2_SB(mnt->mnt_sb); + struct ocfs2_super *osb = OCFS2_SB(root->d_sb); unsigned long opts = osb->s_mount_opt; unsigned int local_alloc_megs; @@ -1566,8 +1550,7 @@ static int ocfs2_show_options(struct seq_file *s, struct vfsmount *mnt) if (osb->preferred_slot != OCFS2_INVALID_SLOT) seq_printf(s, ",preferred_slot=%d", osb->preferred_slot); - if (osb->s_atime_quantum != OCFS2_DEFAULT_ATIME_QUANTUM) - seq_printf(s, ",atime_quantum=%u", osb->s_atime_quantum); + seq_printf(s, ",atime_quantum=%u", osb->s_atime_quantum); if (osb->osb_commit_interval) seq_printf(s, ",commit=%u", @@ -1619,24 +1602,18 @@ static int __init ocfs2_init(void) { int status; - ocfs2_print_version(); - status = init_ocfs2_uptodate_cache(); - if (status < 0) { - mlog_errno(status); - goto leave; - } + if (status < 0) + goto out1; status = ocfs2_initialize_mem_caches(); - if (status < 0) { - mlog_errno(status); - goto leave; - } + if (status < 0) + goto out2; ocfs2_wq = create_singlethread_workqueue("ocfs2_wq"); if (!ocfs2_wq) { status = -ENOMEM; - goto leave; + goto out3; } ocfs2_debugfs_root = debugfs_create_dir("ocfs2", NULL); @@ -1648,17 +1625,23 @@ static int __init ocfs2_init(void) ocfs2_set_locking_protocol(); status = register_quota_format(&ocfs2_quota_format); -leave: - if (status < 0) { - ocfs2_free_mem_caches(); - exit_ocfs2_uptodate_cache(); - mlog_errno(status); - } + if (status < 0) + goto out4; + status = register_filesystem(&ocfs2_fs_type); + if (!status) + return 0; - if (status >= 0) { - return register_filesystem(&ocfs2_fs_type); - } else - return -1; + unregister_quota_format(&ocfs2_quota_format); +out4: + destroy_workqueue(ocfs2_wq); + debugfs_remove(ocfs2_debugfs_root); +out3: + ocfs2_free_mem_caches(); +out2: + exit_ocfs2_uptodate_cache(); +out1: + mlog_errno(status); + return status; } static void __exit ocfs2_exit(void) @@ -1759,7 +1742,7 @@ static void ocfs2_inode_init_once(void *data) ocfs2_extent_map_init(&oi->vfs_inode); INIT_LIST_HEAD(&oi->ip_io_markers); oi->ip_dir_start_lookup = 0; - + mutex_init(&oi->ip_unaligned_aio); init_rwsem(&oi->ip_alloc_sem); init_rwsem(&oi->ip_xattr_sem); mutex_init(&oi->ip_io_mutex); @@ -1814,6 +1797,11 @@ static int ocfs2_initialize_mem_caches(void) static void ocfs2_free_mem_caches(void) { + /* + * Make sure all delayed rcu free inodes are flushed before we + * destroy cache. + */ + rcu_barrier(); if (ocfs2_inode_cachep) kmem_cache_destroy(ocfs2_inode_cachep); ocfs2_inode_cachep = NULL; @@ -1839,8 +1827,8 @@ static int ocfs2_get_sector(struct super_block *sb, *bh = sb_getblk(sb, block); if (!*bh) { - mlog_errno(-EIO); - return -EIO; + mlog_errno(-ENOMEM); + return -ENOMEM; } lock_buffer(*bh); if (!buffer_dirty(*bh)) @@ -1915,7 +1903,7 @@ static void ocfs2_dismount_volume(struct super_block *sb, int mnt_err) { int tmp, hangup_needed = 0; struct ocfs2_super *osb = NULL; - char nodestr[8]; + char nodestr[12]; trace_ocfs2_dismount_volume(sb); @@ -1925,17 +1913,16 @@ static void ocfs2_dismount_volume(struct super_block *sb, int mnt_err) debugfs_remove(osb->osb_ctxt); - /* - * Flush inode dropping work queue so that deletes are - * performed while the filesystem is still working - */ - ocfs2_drop_all_dl_inodes(osb); - /* Orphan scan should be stopped as early as possible */ ocfs2_orphan_scan_stop(osb); ocfs2_disable_quotas(osb); + /* All dquots should be freed by now */ + WARN_ON(!llist_empty(&osb->dquot_drop_list)); + /* Wait for worker to be done with the work structure in osb */ + cancel_work_sync(&osb->dquot_drop_work); + ocfs2_shutdown_local_alloc(osb); ocfs2_truncate_log_shutdown(osb); @@ -1973,7 +1960,8 @@ static void ocfs2_dismount_volume(struct super_block *sb, int mnt_err) * If we failed before we got a uuid_str yet, we can't stop * heartbeat. Otherwise, do it. */ - if (!mnt_err && !ocfs2_mount_local(osb) && osb->uuid_str) + if (!mnt_err && !ocfs2_mount_local(osb) && osb->uuid_str && + !ocfs2_is_hard_readonly(osb)) hangup_needed = 1; if (osb->cconn) @@ -2065,7 +2053,6 @@ static int ocfs2_initialize_super(struct super_block *sb, struct ocfs2_dinode *di = (struct ocfs2_dinode *)bh->b_data; struct inode *inode = NULL; struct ocfs2_journal *journal; - __le32 uuid_net_key; struct ocfs2_super *osb; u64 total_blocks; @@ -2111,6 +2098,8 @@ static int ocfs2_initialize_super(struct super_block *sb, spin_lock_init(&osb->osb_xattr_lock); ocfs2_init_steal_slots(osb); + mutex_init(&osb->system_file_mutex); + atomic_set(&osb->alloc_stats.moves, 0); atomic_set(&osb->alloc_stats.local_data, 0); atomic_set(&osb->alloc_stats.bitmap_data, 0); @@ -2143,7 +2132,6 @@ static int ocfs2_initialize_super(struct super_block *sb, } init_waitqueue_head(&osb->checkpoint_event); - atomic_set(&osb->needs_checkpoint, 0); osb->s_atime_quantum = OCFS2_DEFAULT_ATIME_QUANTUM; @@ -2216,10 +2204,9 @@ static int ocfs2_initialize_super(struct super_block *sb, if (ocfs2_clusterinfo_valid(osb)) { osb->osb_stackflags = OCFS2_RAW_SB(di)->s_cluster_info.ci_stackflags; - memcpy(osb->osb_cluster_stack, + strlcpy(osb->osb_cluster_stack, OCFS2_RAW_SB(di)->s_cluster_info.ci_stack, - OCFS2_STACK_LABEL_LEN); - osb->osb_cluster_stack[OCFS2_STACK_LABEL_LEN] = '\0'; + OCFS2_STACK_LABEL_LEN + 1); if (strlen(osb->osb_cluster_stack) != OCFS2_STACK_LABEL_LEN) { mlog(ML_ERROR, "couldn't mount because of an invalid " @@ -2228,6 +2215,9 @@ static int ocfs2_initialize_super(struct super_block *sb, status = -EINVAL; goto bail; } + strlcpy(osb->osb_cluster_name, + OCFS2_RAW_SB(di)->s_cluster_info.ci_cluster, + OCFS2_CLUSTER_NAME_LEN + 1); } else { /* The empty string is identical with classic tools that * don't know about s_cluster_info. */ @@ -2263,8 +2253,8 @@ static int ocfs2_initialize_super(struct super_block *sb, INIT_WORK(&journal->j_recovery_work, ocfs2_complete_recovery); journal->j_state = OCFS2_JOURNAL_FREE; - INIT_WORK(&osb->dentry_lock_work, ocfs2_drop_dl_inodes); - osb->dentry_lock_list = NULL; + INIT_WORK(&osb->dquot_drop_work, ocfs2_drop_dquot_refs); + init_llist_head(&osb->dquot_drop_list); /* get some pseudo constants for clustersize bits */ osb->s_clustersize_bits = @@ -2298,10 +2288,8 @@ static int ocfs2_initialize_super(struct super_block *sb, goto bail; } - memcpy(&uuid_net_key, di->id2.i_super.s_uuid, sizeof(uuid_net_key)); - - strncpy(osb->vol_label, di->id2.i_super.s_label, 63); - osb->vol_label[63] = '\0'; + strlcpy(osb->vol_label, di->id2.i_super.s_label, + OCFS2_MAX_VOL_LABEL_LEN); osb->root_blkno = le64_to_cpu(di->id2.i_super.s_root_blkno); osb->system_dir_blkno = le64_to_cpu(di->id2.i_super.s_system_dir_blkno); osb->first_cluster_group_blkno = @@ -2352,6 +2340,7 @@ static int ocfs2_initialize_super(struct super_block *sb, mlog_errno(status); goto bail; } + cleancache_init_shared_fs((char *)&di->id2.i_super.s_uuid, sb); bail: return status; @@ -2460,8 +2449,8 @@ static int ocfs2_check_volume(struct ocfs2_super *osb) goto finally; } } else { - mlog(ML_NOTICE, "File system was not unmounted cleanly, " - "recovering volume.\n"); + printk(KERN_NOTICE "ocfs2: File system on device (%s) was not " + "unmounted cleanly, recovering it.\n", osb->dev_str); } local = ocfs2_mount_local(osb); @@ -2514,8 +2503,7 @@ static int ocfs2_check_volume(struct ocfs2_super *osb) mlog_errno(status); finally: - if (local_alloc) - kfree(local_alloc); + kfree(local_alloc); if (status) mlog_errno(status); @@ -2542,8 +2530,7 @@ static void ocfs2_delete_osb(struct ocfs2_super *osb) * we free it here. */ kfree(osb->journal); - if (osb->local_alloc_copy) - kfree(osb->local_alloc_copy); + kfree(osb->local_alloc_copy); kfree(osb->uuid_str); ocfs2_put_dlm_debug(osb->osb_dlm_debug); memset(osb, 0, sizeof(struct ocfs2_super)); diff --git a/fs/ocfs2/super.h b/fs/ocfs2/super.h index 40c7de084c1..74ff74cf78f 100644 --- a/fs/ocfs2/super.h +++ b/fs/ocfs2/super.h @@ -31,17 +31,15 @@ extern struct workqueue_struct *ocfs2_wq; int ocfs2_publish_get_mount_state(struct ocfs2_super *osb, int node_num); -void __ocfs2_error(struct super_block *sb, - const char *function, - const char *fmt, ...) - __attribute__ ((format (printf, 3, 4))); +__printf(3, 4) +void __ocfs2_error(struct super_block *sb, const char *function, + const char *fmt, ...); #define ocfs2_error(sb, fmt, args...) __ocfs2_error(sb, __PRETTY_FUNCTION__, fmt, ##args) -void __ocfs2_abort(struct super_block *sb, - const char *function, - const char *fmt, ...) - __attribute__ ((format (printf, 3, 4))); +__printf(3, 4) +void __ocfs2_abort(struct super_block *sb, const char *function, + const char *fmt, ...); #define ocfs2_abort(sb, fmt, args...) __ocfs2_abort(sb, __PRETTY_FUNCTION__, fmt, ##args) diff --git a/fs/ocfs2/symlink.c b/fs/ocfs2/symlink.c index 5d22872e2bb..66edce7ecfd 100644 --- a/fs/ocfs2/symlink.c +++ b/fs/ocfs2/symlink.c @@ -54,101 +54,40 @@ #include "buffer_head_io.h" -static char *ocfs2_fast_symlink_getlink(struct inode *inode, - struct buffer_head **bh) +static int ocfs2_fast_symlink_readpage(struct file *unused, struct page *page) { - int status; - char *link = NULL; + struct inode *inode = page->mapping->host; + struct buffer_head *bh = NULL; + int status = ocfs2_read_inode_block(inode, &bh); struct ocfs2_dinode *fe; + const char *link; + void *kaddr; + size_t len; - status = ocfs2_read_inode_block(inode, bh); if (status < 0) { mlog_errno(status); - link = ERR_PTR(status); - goto bail; + return status; } - fe = (struct ocfs2_dinode *) (*bh)->b_data; + fe = (struct ocfs2_dinode *) bh->b_data; link = (char *) fe->id2.i_symlink; -bail: - - return link; -} - -static int ocfs2_readlink(struct dentry *dentry, - char __user *buffer, - int buflen) -{ - int ret; - char *link; - struct buffer_head *bh = NULL; - struct inode *inode = dentry->d_inode; - - link = ocfs2_fast_symlink_getlink(inode, &bh); - if (IS_ERR(link)) { - ret = PTR_ERR(link); - goto out; - } - - /* - * Without vfsmount we can't update atime now, - * but we will update atime here ultimately. - */ - ret = vfs_readlink(dentry, buffer, buflen, link); - - brelse(bh); -out: - if (ret < 0) - mlog_errno(ret); - return ret; -} - -static void *ocfs2_fast_follow_link(struct dentry *dentry, - struct nameidata *nd) -{ - int status = 0; - int len; - char *target, *link = ERR_PTR(-ENOMEM); - struct inode *inode = dentry->d_inode; - struct buffer_head *bh = NULL; - - BUG_ON(!ocfs2_inode_is_fast_symlink(inode)); - target = ocfs2_fast_symlink_getlink(inode, &bh); - if (IS_ERR(target)) { - status = PTR_ERR(target); - mlog_errno(status); - goto bail; - } - - /* Fast symlinks can't be large */ - len = strnlen(target, ocfs2_fast_symlink_chars(inode->i_sb)); - link = kzalloc(len + 1, GFP_NOFS); - if (!link) { - status = -ENOMEM; - mlog_errno(status); - goto bail; - } - - memcpy(link, target, len); - -bail: - nd_set_link(nd, status ? ERR_PTR(status) : link); + /* will be less than a page size */ + len = strnlen(link, ocfs2_fast_symlink_chars(inode->i_sb)); + kaddr = kmap_atomic(page); + memcpy(kaddr, link, len + 1); + kunmap_atomic(kaddr); + SetPageUptodate(page); + unlock_page(page); brelse(bh); - - if (status) - mlog_errno(status); - return NULL; + return 0; } -static void ocfs2_fast_put_link(struct dentry *dentry, struct nameidata *nd, void *cookie) -{ - char *link = nd_get_link(nd); - if (!IS_ERR(link)) - kfree(link); -} +const struct address_space_operations ocfs2_fast_symlink_aops = { + .readpage = ocfs2_fast_symlink_readpage, +}; const struct inode_operations ocfs2_symlink_inode_operations = { - .readlink = page_readlink, + .readlink = generic_readlink, .follow_link = page_follow_link_light, .put_link = page_put_link, .getattr = ocfs2_getattr, @@ -159,15 +98,3 @@ const struct inode_operations ocfs2_symlink_inode_operations = { .removexattr = generic_removexattr, .fiemap = ocfs2_fiemap, }; -const struct inode_operations ocfs2_fast_symlink_inode_operations = { - .readlink = ocfs2_readlink, - .follow_link = ocfs2_fast_follow_link, - .put_link = ocfs2_fast_put_link, - .getattr = ocfs2_getattr, - .setattr = ocfs2_setattr, - .setxattr = generic_setxattr, - .getxattr = generic_getxattr, - .listxattr = ocfs2_listxattr, - .removexattr = generic_removexattr, - .fiemap = ocfs2_fiemap, -}; diff --git a/fs/ocfs2/symlink.h b/fs/ocfs2/symlink.h index 65a6c9c6ad5..71ee4245e91 100644 --- a/fs/ocfs2/symlink.h +++ b/fs/ocfs2/symlink.h @@ -27,7 +27,7 @@ #define OCFS2_SYMLINK_H extern const struct inode_operations ocfs2_symlink_inode_operations; -extern const struct inode_operations ocfs2_fast_symlink_inode_operations; +extern const struct address_space_operations ocfs2_fast_symlink_aops; /* * Test whether an inode is a fast symlink. diff --git a/fs/ocfs2/sysfile.c b/fs/ocfs2/sysfile.c index 3d635f4bbb2..af155c18312 100644 --- a/fs/ocfs2/sysfile.c +++ b/fs/ocfs2/sysfile.c @@ -91,8 +91,7 @@ static struct inode **get_local_system_inode(struct ocfs2_super *osb, } else osb->local_system_inodes = local_system_inodes; spin_unlock(&osb->osb_lock); - if (unlikely(free)) - kfree(free); + kfree(free); } index = (slot * NUM_LOCAL_SYSTEM_INODES) + @@ -114,9 +113,11 @@ struct inode *ocfs2_get_system_file_inode(struct ocfs2_super *osb, } else arr = get_local_system_inode(osb, type, slot); + mutex_lock(&osb->system_file_mutex); if (arr && ((inode = *arr) != NULL)) { /* get a ref in addition to the array ref */ inode = igrab(inode); + mutex_unlock(&osb->system_file_mutex); BUG_ON(!inode); return inode; @@ -130,6 +131,7 @@ struct inode *ocfs2_get_system_file_inode(struct ocfs2_super *osb, *arr = igrab(inode); BUG_ON(!*arr); } + mutex_unlock(&osb->system_file_mutex); return inode; } diff --git a/fs/ocfs2/uptodate.c b/fs/ocfs2/uptodate.c index 52eaf33d346..82e17b076ce 100644 --- a/fs/ocfs2/uptodate.c +++ b/fs/ocfs2/uptodate.c @@ -67,7 +67,7 @@ struct ocfs2_meta_cache_item { sector_t c_block; }; -static struct kmem_cache *ocfs2_uptodate_cachep = NULL; +static struct kmem_cache *ocfs2_uptodate_cachep; u64 ocfs2_metadata_cache_owner(struct ocfs2_caching_info *ci) { diff --git a/fs/ocfs2/ver.c b/fs/ocfs2/ver.c deleted file mode 100644 index e2488f4128a..00000000000 --- a/fs/ocfs2/ver.c +++ /dev/null @@ -1,43 +0,0 @@ -/* -*- mode: c; c-basic-offset: 8; -*- - * vim: noexpandtab sw=8 ts=8 sts=0: - * - * ver.c - * - * version string - * - * Copyright (C) 2002, 2005 Oracle. All rights reserved. - * - * This program is free software; you can redistribute it and/or - * modify it under the terms of the GNU General Public - * License as published by the Free Software Foundation; either - * version 2 of the License, or (at your option) any later version. - * - * This program is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * General Public License for more details. - * - * You should have received a copy of the GNU General Public - * License along with this program; if not, write to the - * Free Software Foundation, Inc., 59 Temple Place - Suite 330, - * Boston, MA 021110-1307, USA. - */ - -#include <linux/module.h> -#include <linux/string.h> -#include <linux/kernel.h> - -#include "ver.h" - -#define OCFS2_BUILD_VERSION "1.5.0" - -#define VERSION_STR "OCFS2 " OCFS2_BUILD_VERSION - -void ocfs2_print_version(void) -{ - printk(KERN_INFO "%s\n", VERSION_STR); -} - -MODULE_DESCRIPTION(VERSION_STR); - -MODULE_VERSION(OCFS2_BUILD_VERSION); diff --git a/fs/ocfs2/ver.h b/fs/ocfs2/ver.h deleted file mode 100644 index d7395cb91d2..00000000000 --- a/fs/ocfs2/ver.h +++ /dev/null @@ -1,31 +0,0 @@ -/* -*- mode: c; c-basic-offset: 8; -*- - * vim: noexpandtab sw=8 ts=8 sts=0: - * - * ver.h - * - * Function prototypes - * - * Copyright (C) 2002, 2004 Oracle. All rights reserved. - * - * This program is free software; you can redistribute it and/or - * modify it under the terms of the GNU General Public - * License as published by the Free Software Foundation; either - * version 2 of the License, or (at your option) any later version. - * - * This program is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * General Public License for more details. - * - * You should have received a copy of the GNU General Public - * License along with this program; if not, write to the - * Free Software Foundation, Inc., 59 Temple Place - Suite 330, - * Boston, MA 021110-1307, USA. - */ - -#ifndef OCFS2_VER_H -#define OCFS2_VER_H - -void ocfs2_print_version(void); - -#endif /* OCFS2_VER_H */ diff --git a/fs/ocfs2/xattr.c b/fs/ocfs2/xattr.c index 81ecf9c0bf0..016f01df382 100644 --- a/fs/ocfs2/xattr.c +++ b/fs/ocfs2/xattr.c @@ -99,8 +99,8 @@ static struct ocfs2_xattr_def_value_root def_xv = { const struct xattr_handler *ocfs2_xattr_handlers[] = { &ocfs2_xattr_user_handler, - &ocfs2_xattr_acl_access_handler, - &ocfs2_xattr_acl_default_handler, + &posix_acl_access_xattr_handler, + &posix_acl_default_xattr_handler, &ocfs2_xattr_trusted_handler, &ocfs2_xattr_security_handler, NULL @@ -109,9 +109,9 @@ const struct xattr_handler *ocfs2_xattr_handlers[] = { static const struct xattr_handler *ocfs2_xattr_handler_map[OCFS2_XATTR_MAX] = { [OCFS2_XATTR_INDEX_USER] = &ocfs2_xattr_user_handler, [OCFS2_XATTR_INDEX_POSIX_ACL_ACCESS] - = &ocfs2_xattr_acl_access_handler, + = &posix_acl_access_xattr_handler, [OCFS2_XATTR_INDEX_POSIX_ACL_DEFAULT] - = &ocfs2_xattr_acl_default_handler, + = &posix_acl_default_xattr_handler, [OCFS2_XATTR_INDEX_TRUSTED] = &ocfs2_xattr_trusted_handler, [OCFS2_XATTR_INDEX_SECURITY] = &ocfs2_xattr_security_handler, }; @@ -369,7 +369,7 @@ static void ocfs2_xattr_bucket_free(struct ocfs2_xattr_bucket *bucket) * them fully. */ static int ocfs2_init_xattr_bucket(struct ocfs2_xattr_bucket *bucket, - u64 xb_blkno) + u64 xb_blkno, int new) { int i, rc = 0; @@ -377,15 +377,22 @@ static int ocfs2_init_xattr_bucket(struct ocfs2_xattr_bucket *bucket, bucket->bu_bhs[i] = sb_getblk(bucket->bu_inode->i_sb, xb_blkno + i); if (!bucket->bu_bhs[i]) { - rc = -EIO; + rc = -ENOMEM; mlog_errno(rc); break; } if (!ocfs2_buffer_uptodate(INODE_CACHE(bucket->bu_inode), - bucket->bu_bhs[i])) - ocfs2_set_new_buffer_uptodate(INODE_CACHE(bucket->bu_inode), - bucket->bu_bhs[i]); + bucket->bu_bhs[i])) { + if (new) + ocfs2_set_new_buffer_uptodate(INODE_CACHE(bucket->bu_inode), + bucket->bu_bhs[i]); + else { + set_buffer_uptodate(bucket->bu_bhs[i]); + ocfs2_set_buffer_uptodate(INODE_CACHE(bucket->bu_inode), + bucket->bu_bhs[i]); + } + } } if (rc) @@ -623,7 +630,7 @@ int ocfs2_calc_security_init(struct inode *dir, int ocfs2_calc_xattr_init(struct inode *dir, struct buffer_head *dir_bh, - int mode, + umode_t mode, struct ocfs2_security_xattr_info *si, int *want_clusters, int *xattr_credits, @@ -754,8 +761,7 @@ static int ocfs2_xattr_extend_allocation(struct inode *inode, BUG_ON(why == RESTART_META); credits = ocfs2_calc_extend_credits(inode->i_sb, - &vb->vb_xv->xr_list, - clusters_to_add); + &vb->vb_xv->xr_list); status = ocfs2_extend_trans(handle, credits); if (status < 0) { status = -ENOMEM; @@ -2376,16 +2382,18 @@ static int ocfs2_remove_value_outside(struct inode*inode, } ret = ocfs2_xattr_value_truncate(inode, vb, 0, &ctxt); - if (ret < 0) { - mlog_errno(ret); - break; - } ocfs2_commit_trans(osb, ctxt.handle); if (ctxt.meta_ac) { ocfs2_free_alloc_context(ctxt.meta_ac); ctxt.meta_ac = NULL; } + + if (ret < 0) { + mlog_errno(ret); + break; + } + } if (ctxt.meta_ac) @@ -2601,6 +2609,7 @@ int ocfs2_xattr_remove(struct inode *inode, struct buffer_head *di_bh) oi->ip_dyn_features &= ~(OCFS2_INLINE_XATTR_FL | OCFS2_HAS_XATTR_FL); di->i_dyn_features = cpu_to_le16(oi->ip_dyn_features); spin_unlock(&oi->ip_lock); + ocfs2_update_inode_fsync_trans(handle, inode, 0); ocfs2_journal_dirty(handle, di_bh); out_commit: @@ -2749,7 +2758,6 @@ static int ocfs2_xattr_ibody_set(struct inode *inode, { int ret; struct ocfs2_inode_info *oi = OCFS2_I(inode); - struct ocfs2_dinode *di = (struct ocfs2_dinode *)xs->inode_bh->b_data; struct ocfs2_xa_loc loc; if (inode->i_sb->s_blocksize == OCFS2_MIN_BLOCKSIZE) @@ -2757,13 +2765,6 @@ static int ocfs2_xattr_ibody_set(struct inode *inode, down_write(&oi->ip_alloc_sem); if (!(oi->ip_dyn_features & OCFS2_INLINE_XATTR_FL)) { - if (!ocfs2_xattr_has_space_inline(inode, di)) { - ret = -ENOSPC; - goto out; - } - } - - if (!(oi->ip_dyn_features & OCFS2_INLINE_XATTR_FL)) { ret = ocfs2_xattr_ibody_init(inode, xs->inode_bh, ctxt); if (ret) { if (ret != -ENOSPC) @@ -2871,6 +2872,12 @@ static int ocfs2_create_xattr_block(struct inode *inode, } new_bh = sb_getblk(inode->i_sb, first_blkno); + if (!new_bh) { + ret = -ENOMEM; + mlog_errno(ret); + goto end; + } + ocfs2_set_new_buffer_uptodate(INODE_CACHE(inode), new_bh); ret = ocfs2_journal_access_xb(ctxt->handle, INODE_CACHE(inode), @@ -3046,8 +3053,7 @@ static int ocfs2_calc_xattr_set_need(struct inode *inode, if (xi->xi_value_len > OCFS2_XATTR_INLINE_SIZE) { clusters_add += new_clusters; credits += ocfs2_calc_extend_credits(inode->i_sb, - &def_xv.xv.xr_list, - new_clusters); + &def_xv.xv.xr_list); } goto meta_guess; @@ -3112,8 +3118,7 @@ static int ocfs2_calc_xattr_set_need(struct inode *inode, if (!ocfs2_xattr_is_local(xe)) credits += ocfs2_calc_extend_credits( inode->i_sb, - &def_xv.xv.xr_list, - new_clusters); + &def_xv.xv.xr_list); goto out; } } @@ -3138,9 +3143,7 @@ static int ocfs2_calc_xattr_set_need(struct inode *inode, meta_add += ocfs2_extend_meta_needed(&xv->xr_list); clusters_add += new_clusters - old_clusters; credits += ocfs2_calc_extend_credits(inode->i_sb, - &xv->xr_list, - new_clusters - - old_clusters); + &xv->xr_list); if (value_size >= OCFS2_XATTR_ROOT_SIZE) goto out; } @@ -3186,7 +3189,7 @@ meta_guess: &xb->xb_attrs.xb_root.xt_list; meta_add += ocfs2_extend_meta_needed(el); credits += ocfs2_calc_extend_credits(inode->i_sb, - el, 1); + el); } else credits += OCFS2_SUBALLOC_ALLOC + 1; @@ -3205,8 +3208,15 @@ meta_guess: clusters_add += 1; } } else { - meta_add += 1; credits += OCFS2_XATTR_BLOCK_CREATE_CREDITS; + if (xi->xi_value_len > OCFS2_XATTR_INLINE_SIZE) { + struct ocfs2_extent_list *el = &def_xv.xv.xr_list; + meta_add += ocfs2_extend_meta_needed(el); + credits += ocfs2_calc_extend_credits(inode->i_sb, + el); + } else { + meta_add += 1; + } } out: if (clusters_need) @@ -3511,7 +3521,7 @@ int ocfs2_xattr_set(struct inode *inode, int ret, credits, ref_meta = 0, ref_credits = 0; struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); struct inode *tl_inode = osb->osb_tl_inode; - struct ocfs2_xattr_set_ctxt ctxt = { NULL, NULL, }; + struct ocfs2_xattr_set_ctxt ctxt = { NULL, NULL, NULL, }; struct ocfs2_refcount_tree *ref_tree = NULL; struct ocfs2_xattr_info xi = { @@ -3615,13 +3625,15 @@ int ocfs2_xattr_set(struct inode *inode, if (IS_ERR(ctxt.handle)) { ret = PTR_ERR(ctxt.handle); mlog_errno(ret); - goto cleanup; + goto out_free_ac; } ret = __ocfs2_xattr_set_handle(inode, di, &xi, &xis, &xbs, &ctxt); + ocfs2_update_inode_fsync_trans(ctxt.handle, inode, 0); ocfs2_commit_trans(osb, ctxt.handle); +out_free_ac: if (ctxt.data_ac) ocfs2_free_alloc_context(ctxt.data_ac); if (ctxt.meta_ac) @@ -4298,7 +4310,7 @@ static int ocfs2_xattr_create_index_block(struct inode *inode, trace_ocfs2_xattr_create_index_block((unsigned long long)blkno); - ret = ocfs2_init_xattr_bucket(xs->bucket, blkno); + ret = ocfs2_init_xattr_bucket(xs->bucket, blkno, 1); if (ret) { mlog_errno(ret); goto out; @@ -4642,7 +4654,7 @@ static int ocfs2_divide_xattr_bucket(struct inode *inode, * Even if !new_bucket_head, we're overwriting t_bucket. Thus, * there's no need to read it. */ - ret = ocfs2_init_xattr_bucket(t_bucket, new_blk); + ret = ocfs2_init_xattr_bucket(t_bucket, new_blk, new_bucket_head); if (ret) { mlog_errno(ret); goto out; @@ -4808,7 +4820,7 @@ static int ocfs2_cp_xattr_bucket(struct inode *inode, * Even if !t_is_new, we're overwriting t_bucket. Thus, * there's no need to read it. */ - ret = ocfs2_init_xattr_bucket(t_bucket, t_blkno); + ret = ocfs2_init_xattr_bucket(t_bucket, t_blkno, t_is_new); if (ret) goto out; @@ -5480,6 +5492,7 @@ static int ocfs2_rm_xattr_cluster(struct inode *inode, ret = ocfs2_truncate_log_append(osb, handle, blkno, len); if (ret) mlog_errno(ret); + ocfs2_update_inode_fsync_trans(handle, inode, 0); out_commit: ocfs2_commit_trans(osb, handle); @@ -5887,6 +5900,10 @@ static int ocfs2_xattr_value_attach_refcount(struct inode *inode, while (cpos < clusters) { ret = ocfs2_xattr_get_clusters(inode, cpos, &p_cluster, &num_clusters, el, &ext_flags); + if (ret) { + mlog_errno(ret); + break; + } cpos += num_clusters; if ((ext_flags & OCFS2_EXT_REFCOUNTED)) @@ -6217,8 +6234,7 @@ static int ocfs2_value_metas_in_xattr_header(struct super_block *sb, le16_to_cpu(xv->xr_list.l_next_free_rec); *credits += ocfs2_calc_extend_credits(sb, - &def_xv.xv.xr_list, - le32_to_cpu(xv->xr_clusters)); + &def_xv.xv.xr_list); /* * If the value is a tree with depth > 1, We don't go deep @@ -6497,6 +6513,16 @@ static int ocfs2_reflink_xattr_inline(struct ocfs2_xattr_reflink *args) } new_oi = OCFS2_I(args->new_inode); + /* + * Adjust extent record count to reserve space for extended attribute. + * Inline data count had been adjusted in ocfs2_duplicate_inline_data(). + */ + if (!(new_oi->ip_dyn_features & OCFS2_INLINE_DATA_FL) && + !(ocfs2_inode_is_fast_symlink(args->new_inode))) { + struct ocfs2_extent_list *el = &new_di->id2.i_list; + le16_add_cpu(&el->l_count, -(inline_size / + sizeof(struct ocfs2_extent_rec))); + } spin_lock(&new_oi->ip_lock); new_oi->ip_dyn_features |= OCFS2_HAS_XATTR_FL | OCFS2_INLINE_XATTR_FL; new_di->i_dyn_features = cpu_to_le16(new_oi->ip_dyn_features); @@ -6773,7 +6799,7 @@ static int ocfs2_lock_reflink_xattr_rec_allocators( metas.num_metas += ocfs2_extend_meta_needed(xt_et->et_root_el); *credits += ocfs2_calc_extend_credits(osb->sb, - xt_et->et_root_el, len); + xt_et->et_root_el); if (metas.num_metas) { ret = ocfs2_reserve_new_metadata_blocks(osb, metas.num_metas, @@ -6793,7 +6819,7 @@ out: if (ret) { if (*meta_ac) { ocfs2_free_alloc_context(*meta_ac); - meta_ac = NULL; + *meta_ac = NULL; } } @@ -6821,7 +6847,7 @@ static int ocfs2_reflink_xattr_bucket(handle_t *handle, break; } - ret = ocfs2_init_xattr_bucket(args->new_bucket, new_blkno); + ret = ocfs2_init_xattr_bucket(args->new_bucket, new_blkno, 1); if (ret) { mlog_errno(ret); break; @@ -7181,24 +7207,15 @@ out: */ int ocfs2_init_security_and_acl(struct inode *dir, struct inode *inode, - const struct qstr *qstr) + const struct qstr *qstr, + struct posix_acl *default_acl, + struct posix_acl *acl) { - int ret = 0; struct buffer_head *dir_bh = NULL; - struct ocfs2_security_xattr_info si = { - .enable = 1, - }; + int ret = 0; - ret = ocfs2_init_security_get(inode, dir, qstr, &si); - if (!ret) { - ret = ocfs2_xattr_set(inode, OCFS2_XATTR_INDEX_SECURITY, - si.name, si.value, si.value_len, - XATTR_CREATE); - if (ret) { - mlog_errno(ret); - goto leave; - } - } else if (ret != -EOPNOTSUPP) { + ret = ocfs2_init_security_get(inode, dir, qstr, NULL); + if (ret) { mlog_errno(ret); goto leave; } @@ -7209,9 +7226,10 @@ int ocfs2_init_security_and_acl(struct inode *dir, goto leave; } - ret = ocfs2_init_acl(NULL, inode, dir, NULL, dir_bh, NULL, NULL); - if (ret) - mlog_errno(ret); + if (!ret && default_acl) + ret = ocfs2_iop_set_acl(inode, default_acl, ACL_TYPE_DEFAULT); + if (!ret && acl) + ret = ocfs2_iop_set_acl(inode, acl, ACL_TYPE_ACCESS); ocfs2_inode_unlock(dir, 0); brelse(dir_bh); @@ -7255,6 +7273,22 @@ static int ocfs2_xattr_security_set(struct dentry *dentry, const char *name, name, value, size, flags); } +int ocfs2_initxattrs(struct inode *inode, const struct xattr *xattr_array, + void *fs_info) +{ + const struct xattr *xattr; + int err = 0; + + for (xattr = xattr_array; xattr->name != NULL; xattr++) { + err = ocfs2_xattr_set(inode, OCFS2_XATTR_INDEX_SECURITY, + xattr->name, xattr->value, + xattr->value_len, XATTR_CREATE); + if (err) + break; + } + return err; +} + int ocfs2_init_security_get(struct inode *inode, struct inode *dir, const struct qstr *qstr, @@ -7263,8 +7297,13 @@ int ocfs2_init_security_get(struct inode *inode, /* check whether ocfs2 support feature xattr */ if (!ocfs2_supports_xattr(OCFS2_SB(dir->i_sb))) return -EOPNOTSUPP; - return security_inode_init_security(inode, dir, qstr, &si->name, - &si->value, &si->value_len); + if (si) + return security_old_inode_init_security(inode, dir, qstr, + &si->name, &si->value, + &si->value_len); + + return security_inode_init_security(inode, dir, qstr, + &ocfs2_initxattrs, NULL); } int ocfs2_init_security_set(handle_t *handle, diff --git a/fs/ocfs2/xattr.h b/fs/ocfs2/xattr.h index d63cfb72316..f10d5b93c36 100644 --- a/fs/ocfs2/xattr.h +++ b/fs/ocfs2/xattr.h @@ -32,7 +32,7 @@ enum ocfs2_xattr_type { struct ocfs2_security_xattr_info { int enable; - char *name; + const char *name; void *value; size_t value_len; }; @@ -40,8 +40,6 @@ struct ocfs2_security_xattr_info { extern const struct xattr_handler ocfs2_xattr_user_handler; extern const struct xattr_handler ocfs2_xattr_trusted_handler; extern const struct xattr_handler ocfs2_xattr_security_handler; -extern const struct xattr_handler ocfs2_xattr_acl_access_handler; -extern const struct xattr_handler ocfs2_xattr_acl_default_handler; extern const struct xattr_handler *ocfs2_xattr_handlers[]; ssize_t ocfs2_listxattr(struct dentry *, char *, size_t); @@ -68,7 +66,7 @@ int ocfs2_calc_security_init(struct inode *, struct ocfs2_security_xattr_info *, int *, int *, struct ocfs2_alloc_context **); int ocfs2_calc_xattr_init(struct inode *, struct buffer_head *, - int, struct ocfs2_security_xattr_info *, + umode_t, struct ocfs2_security_xattr_info *, int *, int *, int *); /* @@ -96,5 +94,7 @@ int ocfs2_reflink_xattrs(struct inode *old_inode, bool preserve_security); int ocfs2_init_security_and_acl(struct inode *dir, struct inode *inode, - const struct qstr *qstr); + const struct qstr *qstr, + struct posix_acl *default_acl, + struct posix_acl *acl); #endif /* OCFS2_XATTR_H */ |