diff options
Diffstat (limited to 'fs/ocfs2')
46 files changed, 2563 insertions, 421 deletions
diff --git a/fs/ocfs2/acl.c b/fs/ocfs2/acl.c index da702294d7e..391915093fe 100644 --- a/fs/ocfs2/acl.c +++ b/fs/ocfs2/acl.c @@ -209,7 +209,10 @@ static int ocfs2_acl_set_mode(struct inode *inode, struct buffer_head *di_bh, } inode->i_mode = new_mode; + inode->i_ctime = CURRENT_TIME; 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_journal_dirty(handle, di_bh); @@ -290,12 +293,30 @@ static int ocfs2_set_acl(handle_t *handle, int ocfs2_check_acl(struct inode *inode, int mask) { - struct posix_acl *acl = ocfs2_get_acl(inode, ACL_TYPE_ACCESS); + struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); + struct buffer_head *di_bh = NULL; + struct posix_acl *acl; + int ret = -EAGAIN; - if (IS_ERR(acl)) + if (!(osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL)) + return ret; + + ret = ocfs2_read_inode_block(inode, &di_bh); + if (ret < 0) { + mlog_errno(ret); + return ret; + } + + acl = ocfs2_get_acl_nolock(inode, ACL_TYPE_ACCESS, di_bh); + + brelse(di_bh); + + if (IS_ERR(acl)) { + mlog_errno(PTR_ERR(acl)); return PTR_ERR(acl); + } if (acl) { - int ret = posix_acl_permission(inode, acl, mask); + ret = posix_acl_permission(inode, acl, mask); posix_acl_release(acl); return ret; } @@ -344,7 +365,7 @@ int ocfs2_init_acl(handle_t *handle, { struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); struct posix_acl *acl = NULL; - int ret = 0; + int ret = 0, ret2; mode_t mode; if (!S_ISLNK(inode->i_mode)) { @@ -381,7 +402,12 @@ int ocfs2_init_acl(handle_t *handle, mode = inode->i_mode; ret = posix_acl_create_masq(clone, &mode); if (ret >= 0) { - ret = ocfs2_acl_set_mode(inode, di_bh, handle, mode); + 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, diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c index 215e12ce1d8..592fae5007d 100644 --- a/fs/ocfs2/alloc.c +++ b/fs/ocfs2/alloc.c @@ -6672,7 +6672,7 @@ int ocfs2_grab_pages(struct inode *inode, loff_t start, loff_t end, last_page_bytes = PAGE_ALIGN(end); index = start >> PAGE_CACHE_SHIFT; do { - pages[numpages] = grab_cache_page(mapping, index); + pages[numpages] = find_or_create_page(mapping, index, GFP_NOFS); if (!pages[numpages]) { ret = -ENOMEM; mlog_errno(ret); diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c index 0de69c9a08b..5cfeee11815 100644 --- a/fs/ocfs2/aops.c +++ b/fs/ocfs2/aops.c @@ -883,8 +883,8 @@ struct ocfs2_write_ctxt { * out in so that future reads from that region will get * zero's. */ - struct page *w_pages[OCFS2_MAX_CTXT_PAGES]; unsigned int w_num_pages; + struct page *w_pages[OCFS2_MAX_CTXT_PAGES]; struct page *w_target_page; /* @@ -1642,7 +1642,8 @@ static int ocfs2_zero_tail(struct inode *inode, struct buffer_head *di_bh, return ret; } -int ocfs2_write_begin_nolock(struct address_space *mapping, +int ocfs2_write_begin_nolock(struct file *filp, + struct address_space *mapping, loff_t pos, unsigned len, unsigned flags, struct page **pagep, void **fsdata, struct buffer_head *di_bh, struct page *mmap_page) @@ -1692,7 +1693,7 @@ int ocfs2_write_begin_nolock(struct address_space *mapping, mlog_errno(ret); goto out; } else if (ret == 1) { - ret = ocfs2_refcount_cow(inode, di_bh, + ret = ocfs2_refcount_cow(inode, filp, di_bh, wc->w_cpos, wc->w_clen, UINT_MAX); if (ret) { mlog_errno(ret); @@ -1854,7 +1855,7 @@ static int ocfs2_write_begin(struct file *file, struct address_space *mapping, */ down_write(&OCFS2_I(inode)->ip_alloc_sem); - ret = ocfs2_write_begin_nolock(mapping, pos, len, flags, pagep, + ret = ocfs2_write_begin_nolock(file, mapping, pos, len, flags, pagep, fsdata, di_bh, NULL); if (ret) { mlog_errno(ret); diff --git a/fs/ocfs2/aops.h b/fs/ocfs2/aops.h index c48e93ffc51..7606f663da6 100644 --- a/fs/ocfs2/aops.h +++ b/fs/ocfs2/aops.h @@ -48,7 +48,8 @@ int ocfs2_write_end_nolock(struct address_space *mapping, loff_t pos, unsigned len, unsigned copied, struct page *page, void *fsdata); -int ocfs2_write_begin_nolock(struct address_space *mapping, +int ocfs2_write_begin_nolock(struct file *filp, + struct address_space *mapping, loff_t pos, unsigned len, unsigned flags, struct page **pagep, void **fsdata, struct buffer_head *di_bh, struct page *mmap_page); diff --git a/fs/ocfs2/blockcheck.c b/fs/ocfs2/blockcheck.c index ec6d1233959..c7ee03c2222 100644 --- a/fs/ocfs2/blockcheck.c +++ b/fs/ocfs2/blockcheck.c @@ -439,7 +439,7 @@ int ocfs2_block_check_validate(void *data, size_t blocksize, ocfs2_blockcheck_inc_failure(stats); mlog(ML_ERROR, - "CRC32 failed: stored: %u, computed %u. Applying ECC.\n", + "CRC32 failed: stored: 0x%x, computed 0x%x. Applying ECC.\n", (unsigned int)check.bc_crc32e, (unsigned int)crc); /* Ok, try ECC fixups */ @@ -453,7 +453,7 @@ int ocfs2_block_check_validate(void *data, size_t blocksize, goto out; } - mlog(ML_ERROR, "Fixed CRC32 failed: stored: %u, computed %u\n", + mlog(ML_ERROR, "Fixed CRC32 failed: stored: 0x%x, computed 0x%x\n", (unsigned int)check.bc_crc32e, (unsigned int)crc); rc = -EIO; diff --git a/fs/ocfs2/cluster/heartbeat.c b/fs/ocfs2/cluster/heartbeat.c index 41d5f1f92d5..52c7557f3e2 100644 --- a/fs/ocfs2/cluster/heartbeat.c +++ b/fs/ocfs2/cluster/heartbeat.c @@ -62,10 +62,51 @@ static unsigned long o2hb_live_node_bitmap[BITS_TO_LONGS(O2NM_MAX_NODES)]; static LIST_HEAD(o2hb_node_events); static DECLARE_WAIT_QUEUE_HEAD(o2hb_steady_queue); +/* + * In global heartbeat, we maintain a series of region bitmaps. + * - o2hb_region_bitmap allows us to limit the region number to max region. + * - o2hb_live_region_bitmap tracks live regions (seen steady iterations). + * - o2hb_quorum_region_bitmap tracks live regions that have seen all nodes + * heartbeat on it. + * - o2hb_failed_region_bitmap tracks the regions that have seen io timeouts. + */ +static unsigned long o2hb_region_bitmap[BITS_TO_LONGS(O2NM_MAX_REGIONS)]; +static unsigned long o2hb_live_region_bitmap[BITS_TO_LONGS(O2NM_MAX_REGIONS)]; +static unsigned long o2hb_quorum_region_bitmap[BITS_TO_LONGS(O2NM_MAX_REGIONS)]; +static unsigned long o2hb_failed_region_bitmap[BITS_TO_LONGS(O2NM_MAX_REGIONS)]; + +#define O2HB_DB_TYPE_LIVENODES 0 +#define O2HB_DB_TYPE_LIVEREGIONS 1 +#define O2HB_DB_TYPE_QUORUMREGIONS 2 +#define O2HB_DB_TYPE_FAILEDREGIONS 3 +#define O2HB_DB_TYPE_REGION_LIVENODES 4 +#define O2HB_DB_TYPE_REGION_NUMBER 5 +#define O2HB_DB_TYPE_REGION_ELAPSED_TIME 6 +struct o2hb_debug_buf { + int db_type; + int db_size; + int db_len; + void *db_data; +}; + +static struct o2hb_debug_buf *o2hb_db_livenodes; +static struct o2hb_debug_buf *o2hb_db_liveregions; +static struct o2hb_debug_buf *o2hb_db_quorumregions; +static struct o2hb_debug_buf *o2hb_db_failedregions; + #define O2HB_DEBUG_DIR "o2hb" #define O2HB_DEBUG_LIVENODES "livenodes" +#define O2HB_DEBUG_LIVEREGIONS "live_regions" +#define O2HB_DEBUG_QUORUMREGIONS "quorum_regions" +#define O2HB_DEBUG_FAILEDREGIONS "failed_regions" +#define O2HB_DEBUG_REGION_NUMBER "num" +#define O2HB_DEBUG_REGION_ELAPSED_TIME "elapsed_time_in_ms" + static struct dentry *o2hb_debug_dir; static struct dentry *o2hb_debug_livenodes; +static struct dentry *o2hb_debug_liveregions; +static struct dentry *o2hb_debug_quorumregions; +static struct dentry *o2hb_debug_failedregions; static LIST_HEAD(o2hb_all_regions); @@ -77,7 +118,19 @@ static struct o2hb_callback *hbcall_from_type(enum o2hb_callback_type type); #define O2HB_DEFAULT_BLOCK_BITS 9 +enum o2hb_heartbeat_modes { + O2HB_HEARTBEAT_LOCAL = 0, + O2HB_HEARTBEAT_GLOBAL, + O2HB_HEARTBEAT_NUM_MODES, +}; + +char *o2hb_heartbeat_mode_desc[O2HB_HEARTBEAT_NUM_MODES] = { + "local", /* O2HB_HEARTBEAT_LOCAL */ + "global", /* O2HB_HEARTBEAT_GLOBAL */ +}; + unsigned int o2hb_dead_threshold = O2HB_DEFAULT_DEAD_THRESHOLD; +unsigned int o2hb_heartbeat_mode = O2HB_HEARTBEAT_LOCAL; /* Only sets a new threshold if there are no active regions. * @@ -94,6 +147,22 @@ static void o2hb_dead_threshold_set(unsigned int threshold) } } +static int o2hb_global_hearbeat_mode_set(unsigned int hb_mode) +{ + int ret = -1; + + if (hb_mode < O2HB_HEARTBEAT_NUM_MODES) { + spin_lock(&o2hb_live_lock); + if (list_empty(&o2hb_all_regions)) { + o2hb_heartbeat_mode = hb_mode; + ret = 0; + } + spin_unlock(&o2hb_live_lock); + } + + return ret; +} + struct o2hb_node_event { struct list_head hn_item; enum o2hb_callback_type hn_event_type; @@ -135,6 +204,18 @@ struct o2hb_region { struct block_device *hr_bdev; struct o2hb_disk_slot *hr_slots; + /* live node map of this region */ + unsigned long hr_live_node_bitmap[BITS_TO_LONGS(O2NM_MAX_NODES)]; + unsigned int hr_region_num; + + struct dentry *hr_debug_dir; + struct dentry *hr_debug_livenodes; + struct dentry *hr_debug_regnum; + struct dentry *hr_debug_elapsed_time; + struct o2hb_debug_buf *hr_db_livenodes; + struct o2hb_debug_buf *hr_db_regnum; + struct o2hb_debug_buf *hr_db_elapsed_time; + /* let the person setting up hb wait for it to return until it * has reached a 'steady' state. This will be fixed when we have * a more complete api that doesn't lead to this sort of fragility. */ @@ -163,8 +244,19 @@ 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; + unsigned long flags; struct o2hb_region *reg = container_of(work, struct o2hb_region, hr_write_timeout_work.work); @@ -172,6 +264,28 @@ static void o2hb_write_timeout(struct work_struct *work) mlog(ML_ERROR, "Heartbeat write timeout to device %s after %u " "milliseconds\n", reg->hr_dev_name, jiffies_to_msecs(jiffies - reg->hr_last_timeout_start)); + + if (o2hb_global_heartbeat_active()) { + 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, + O2NM_MAX_REGIONS); + quorum = o2hb_pop_count(&o2hb_quorum_region_bitmap, + O2NM_MAX_REGIONS); + spin_unlock_irqrestore(&o2hb_live_lock, flags); + + mlog(ML_HEARTBEAT, "Number of regions %d, failed regions %d\n", + quorum, failed); + + /* + * Fence if the number of failed regions >= half the number + * of quorum regions + */ + if ((failed << 1) < quorum) + return; + } + o2quo_disk_timeout(); } @@ -180,6 +294,11 @@ static void o2hb_arm_write_timeout(struct o2hb_region *reg) mlog(ML_HEARTBEAT, "Queue write timeout for %u ms\n", O2HB_MAX_WRITE_TIMEOUT_MS); + if (o2hb_global_heartbeat_active()) { + spin_lock(&o2hb_live_lock); + clear_bit(reg->hr_region_num, o2hb_failed_region_bitmap); + spin_unlock(&o2hb_live_lock); + } cancel_delayed_work(®->hr_write_timeout_work); reg->hr_last_timeout_start = jiffies; schedule_delayed_work(®->hr_write_timeout_work, @@ -513,6 +632,8 @@ static void o2hb_queue_node_event(struct o2hb_node_event *event, { assert_spin_locked(&o2hb_live_lock); + BUG_ON((!node) && (type != O2HB_NODE_DOWN_CB)); + event->hn_event_type = type; event->hn_node = node; event->hn_node_num = node_num; @@ -554,6 +675,35 @@ static void o2hb_shutdown_slot(struct o2hb_disk_slot *slot) o2nm_node_put(node); } +static void o2hb_set_quorum_device(struct o2hb_region *reg, + struct o2hb_disk_slot *slot) +{ + assert_spin_locked(&o2hb_live_lock); + + if (!o2hb_global_heartbeat_active()) + return; + + if (test_bit(reg->hr_region_num, o2hb_quorum_region_bitmap)) + return; + + /* + * 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 + * added to all nodes. + */ + 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; + + printk(KERN_NOTICE "o2hb: Region %s is now a quorum device\n", + config_item_name(®->hr_item)); + + set_bit(reg->hr_region_num, o2hb_quorum_region_bitmap); +} + static int o2hb_check_slot(struct o2hb_region *reg, struct o2hb_disk_slot *slot) { @@ -565,14 +715,22 @@ static int o2hb_check_slot(struct o2hb_region *reg, u64 cputime; unsigned int dead_ms = o2hb_dead_threshold * O2HB_REGION_TIMEOUT_MS; unsigned int slot_dead_ms; + int tmp; memcpy(hb_block, slot->ds_raw_block, reg->hr_block_bytes); - /* Is this correct? Do we assume that the node doesn't exist - * if we're not configured for him? */ + /* + * If a node is no longer configured but is still in the livemap, we + * may need to clear that bit from the livemap. + */ node = o2nm_get_node_by_num(slot->ds_node_num); - if (!node) - return 0; + if (!node) { + spin_lock(&o2hb_live_lock); + tmp = test_bit(slot->ds_node_num, o2hb_live_node_bitmap); + spin_unlock(&o2hb_live_lock); + if (!tmp) + return 0; + } if (!o2hb_verify_crc(reg, hb_block)) { /* all paths from here will drop o2hb_live_lock for @@ -639,8 +797,12 @@ fire_callbacks: mlog(ML_HEARTBEAT, "Node %d (id 0x%llx) joined my region\n", slot->ds_node_num, (long long)slot->ds_last_generation); + set_bit(slot->ds_node_num, reg->hr_live_node_bitmap); + /* first on the list generates a callback */ if (list_empty(&o2hb_live_slots[slot->ds_node_num])) { + mlog(ML_HEARTBEAT, "o2hb: Add node %d to live nodes " + "bitmap\n", slot->ds_node_num); set_bit(slot->ds_node_num, o2hb_live_node_bitmap); o2hb_queue_node_event(&event, O2HB_NODE_UP_CB, node, @@ -684,13 +846,18 @@ fire_callbacks: mlog(ML_HEARTBEAT, "Node %d left my region\n", slot->ds_node_num); + clear_bit(slot->ds_node_num, reg->hr_live_node_bitmap); + /* last off the live_slot generates a callback */ list_del_init(&slot->ds_live_item); if (list_empty(&o2hb_live_slots[slot->ds_node_num])) { + mlog(ML_HEARTBEAT, "o2hb: Remove node %d from live " + "nodes bitmap\n", slot->ds_node_num); clear_bit(slot->ds_node_num, o2hb_live_node_bitmap); - o2hb_queue_node_event(&event, O2HB_NODE_DOWN_CB, node, - slot->ds_node_num); + /* node can be null */ + o2hb_queue_node_event(&event, O2HB_NODE_DOWN_CB, + node, slot->ds_node_num); changed = 1; } @@ -706,11 +873,14 @@ fire_callbacks: slot->ds_equal_samples = 0; } out: + o2hb_set_quorum_device(reg, slot); + spin_unlock(&o2hb_live_lock); o2hb_run_event_list(&event); - o2nm_node_put(node); + if (node) + o2nm_node_put(node); return changed; } @@ -737,6 +907,7 @@ static int o2hb_do_disk_heartbeat(struct o2hb_region *reg) { int i, ret, highest_node, change = 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; ret = o2nm_configured_node_map(configured_nodes, @@ -746,6 +917,17 @@ static int o2hb_do_disk_heartbeat(struct o2hb_region *reg) return ret; } + /* + * If a node is not configured but is in the livemap, we still need + * to read the slot so as to be able to remove it from the livemap. + */ + o2hb_fill_node_map(live_node_bitmap, sizeof(live_node_bitmap)); + i = -1; + while ((i = find_next_bit(live_node_bitmap, + O2NM_MAX_NODES, i + 1)) < O2NM_MAX_NODES) { + set_bit(i, configured_nodes); + } + 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"); @@ -917,21 +1099,59 @@ static int o2hb_thread(void *data) #ifdef CONFIG_DEBUG_FS 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)]; char *buf = NULL; int i = -1; int out = 0; + /* max_nodes should be the largest bitmap we pass here */ + BUG_ON(sizeof(map) < db->db_size); + buf = kmalloc(PAGE_SIZE, GFP_KERNEL); if (!buf) goto bail; - o2hb_fill_node_map(map, sizeof(map)); + switch (db->db_type) { + case O2HB_DB_TYPE_LIVENODES: + case O2HB_DB_TYPE_LIVEREGIONS: + case O2HB_DB_TYPE_QUORUMREGIONS: + case O2HB_DB_TYPE_FAILEDREGIONS: + spin_lock(&o2hb_live_lock); + memcpy(map, db->db_data, db->db_size); + spin_unlock(&o2hb_live_lock); + break; + + case O2HB_DB_TYPE_REGION_LIVENODES: + spin_lock(&o2hb_live_lock); + reg = (struct o2hb_region *)db->db_data; + memcpy(map, reg->hr_live_node_bitmap, db->db_size); + spin_unlock(&o2hb_live_lock); + break; + + case O2HB_DB_TYPE_REGION_NUMBER: + reg = (struct o2hb_region *)db->db_data; + out += snprintf(buf + out, PAGE_SIZE - out, "%d\n", + reg->hr_region_num); + goto done; + + 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)); + goto done; + + default: + goto done; + } - while ((i = find_next_bit(map, O2NM_MAX_NODES, i + 1)) < O2NM_MAX_NODES) + while ((i = find_next_bit(map, db->db_len, i + 1)) < db->db_len) out += snprintf(buf + out, PAGE_SIZE - out, "%d ", i); out += snprintf(buf + out, PAGE_SIZE - out, "\n"); +done: i_size_write(inode, out); file->private_data = buf; @@ -978,10 +1198,104 @@ static const struct file_operations o2hb_debug_fops = { void o2hb_exit(void) { - if (o2hb_debug_livenodes) - debugfs_remove(o2hb_debug_livenodes); - if (o2hb_debug_dir) - debugfs_remove(o2hb_debug_dir); + kfree(o2hb_db_livenodes); + kfree(o2hb_db_liveregions); + kfree(o2hb_db_quorumregions); + kfree(o2hb_db_failedregions); + debugfs_remove(o2hb_debug_failedregions); + debugfs_remove(o2hb_debug_quorumregions); + debugfs_remove(o2hb_debug_liveregions); + debugfs_remove(o2hb_debug_livenodes); + debugfs_remove(o2hb_debug_dir); +} + +static struct dentry *o2hb_debug_create(const char *name, struct dentry *dir, + struct o2hb_debug_buf **db, int db_len, + int type, int size, int len, void *data) +{ + *db = kmalloc(db_len, GFP_KERNEL); + if (!*db) + return NULL; + + (*db)->db_type = type; + (*db)->db_size = size; + (*db)->db_len = len; + (*db)->db_data = data; + + return debugfs_create_file(name, S_IFREG|S_IRUSR, dir, *db, + &o2hb_debug_fops); +} + +static int o2hb_debug_init(void) +{ + int ret = -ENOMEM; + + o2hb_debug_dir = debugfs_create_dir(O2HB_DEBUG_DIR, NULL); + if (!o2hb_debug_dir) { + mlog_errno(ret); + goto bail; + } + + o2hb_debug_livenodes = o2hb_debug_create(O2HB_DEBUG_LIVENODES, + o2hb_debug_dir, + &o2hb_db_livenodes, + sizeof(*o2hb_db_livenodes), + O2HB_DB_TYPE_LIVENODES, + sizeof(o2hb_live_node_bitmap), + O2NM_MAX_NODES, + o2hb_live_node_bitmap); + if (!o2hb_debug_livenodes) { + mlog_errno(ret); + goto bail; + } + + o2hb_debug_liveregions = o2hb_debug_create(O2HB_DEBUG_LIVEREGIONS, + o2hb_debug_dir, + &o2hb_db_liveregions, + sizeof(*o2hb_db_liveregions), + O2HB_DB_TYPE_LIVEREGIONS, + sizeof(o2hb_live_region_bitmap), + O2NM_MAX_REGIONS, + o2hb_live_region_bitmap); + if (!o2hb_debug_liveregions) { + mlog_errno(ret); + goto bail; + } + + o2hb_debug_quorumregions = + o2hb_debug_create(O2HB_DEBUG_QUORUMREGIONS, + o2hb_debug_dir, + &o2hb_db_quorumregions, + sizeof(*o2hb_db_quorumregions), + O2HB_DB_TYPE_QUORUMREGIONS, + sizeof(o2hb_quorum_region_bitmap), + O2NM_MAX_REGIONS, + o2hb_quorum_region_bitmap); + if (!o2hb_debug_quorumregions) { + mlog_errno(ret); + goto bail; + } + + o2hb_debug_failedregions = + o2hb_debug_create(O2HB_DEBUG_FAILEDREGIONS, + o2hb_debug_dir, + &o2hb_db_failedregions, + sizeof(*o2hb_db_failedregions), + O2HB_DB_TYPE_FAILEDREGIONS, + sizeof(o2hb_failed_region_bitmap), + O2NM_MAX_REGIONS, + o2hb_failed_region_bitmap); + if (!o2hb_debug_failedregions) { + mlog_errno(ret); + goto bail; + } + + ret = 0; +bail: + if (ret) + o2hb_exit(); + + return ret; } int o2hb_init(void) @@ -997,24 +1311,12 @@ int o2hb_init(void) INIT_LIST_HEAD(&o2hb_node_events); memset(o2hb_live_node_bitmap, 0, sizeof(o2hb_live_node_bitmap)); + memset(o2hb_region_bitmap, 0, sizeof(o2hb_region_bitmap)); + memset(o2hb_live_region_bitmap, 0, sizeof(o2hb_live_region_bitmap)); + memset(o2hb_quorum_region_bitmap, 0, sizeof(o2hb_quorum_region_bitmap)); + memset(o2hb_failed_region_bitmap, 0, sizeof(o2hb_failed_region_bitmap)); - o2hb_debug_dir = debugfs_create_dir(O2HB_DEBUG_DIR, NULL); - if (!o2hb_debug_dir) { - mlog_errno(-ENOMEM); - return -ENOMEM; - } - - o2hb_debug_livenodes = debugfs_create_file(O2HB_DEBUG_LIVENODES, - S_IFREG|S_IRUSR, - o2hb_debug_dir, NULL, - &o2hb_debug_fops); - if (!o2hb_debug_livenodes) { - mlog_errno(-ENOMEM); - debugfs_remove(o2hb_debug_dir); - return -ENOMEM; - } - - return 0; + return o2hb_debug_init(); } /* if we're already in a callback then we're already serialized by the sem */ @@ -1078,6 +1380,13 @@ static void o2hb_region_release(struct config_item *item) if (reg->hr_slots) kfree(reg->hr_slots); + kfree(reg->hr_db_regnum); + kfree(reg->hr_db_livenodes); + debugfs_remove(reg->hr_debug_livenodes); + debugfs_remove(reg->hr_debug_regnum); + debugfs_remove(reg->hr_debug_elapsed_time); + debugfs_remove(reg->hr_debug_dir); + spin_lock(&o2hb_live_lock); list_del(®->hr_all_item); spin_unlock(&o2hb_live_lock); @@ -1441,6 +1750,8 @@ static ssize_t o2hb_region_dev_write(struct o2hb_region *reg, /* Ok, we were woken. Make sure it wasn't by drop_item() */ spin_lock(&o2hb_live_lock); hb_task = reg->hr_task; + if (o2hb_global_heartbeat_active()) + set_bit(reg->hr_region_num, o2hb_live_region_bitmap); spin_unlock(&o2hb_live_lock); if (hb_task) @@ -1448,6 +1759,10 @@ static ssize_t o2hb_region_dev_write(struct o2hb_region *reg, else ret = -EIO; + if (hb_task && o2hb_global_heartbeat_active()) + printk(KERN_NOTICE "o2hb: Heartbeat started on region %s\n", + config_item_name(®->hr_item)); + out: if (filp) fput(filp); @@ -1586,21 +1901,94 @@ static struct o2hb_heartbeat_group *to_o2hb_heartbeat_group(struct config_group : NULL; } +static int o2hb_debug_region_init(struct o2hb_region *reg, struct dentry *dir) +{ + int ret = -ENOMEM; + + reg->hr_debug_dir = + debugfs_create_dir(config_item_name(®->hr_item), dir); + if (!reg->hr_debug_dir) { + mlog_errno(ret); + goto bail; + } + + reg->hr_debug_livenodes = + o2hb_debug_create(O2HB_DEBUG_LIVENODES, + reg->hr_debug_dir, + &(reg->hr_db_livenodes), + sizeof(*(reg->hr_db_livenodes)), + O2HB_DB_TYPE_REGION_LIVENODES, + sizeof(reg->hr_live_node_bitmap), + O2NM_MAX_NODES, reg); + if (!reg->hr_debug_livenodes) { + mlog_errno(ret); + goto bail; + } + + reg->hr_debug_regnum = + o2hb_debug_create(O2HB_DEBUG_REGION_NUMBER, + reg->hr_debug_dir, + &(reg->hr_db_regnum), + sizeof(*(reg->hr_db_regnum)), + O2HB_DB_TYPE_REGION_NUMBER, + 0, O2NM_MAX_NODES, reg); + if (!reg->hr_debug_regnum) { + mlog_errno(ret); + goto bail; + } + + reg->hr_debug_elapsed_time = + o2hb_debug_create(O2HB_DEBUG_REGION_ELAPSED_TIME, + reg->hr_debug_dir, + &(reg->hr_db_elapsed_time), + sizeof(*(reg->hr_db_elapsed_time)), + O2HB_DB_TYPE_REGION_ELAPSED_TIME, + 0, 0, reg); + if (!reg->hr_debug_elapsed_time) { + mlog_errno(ret); + goto bail; + } + + ret = 0; +bail: + return ret; +} + static struct config_item *o2hb_heartbeat_group_make_item(struct config_group *group, const char *name) { struct o2hb_region *reg = NULL; + int ret; reg = kzalloc(sizeof(struct o2hb_region), GFP_KERNEL); if (reg == NULL) return ERR_PTR(-ENOMEM); - config_item_init_type_name(®->hr_item, name, &o2hb_region_type); + if (strlen(name) > O2HB_MAX_REGION_NAME_LEN) + return ERR_PTR(-ENAMETOOLONG); spin_lock(&o2hb_live_lock); + reg->hr_region_num = 0; + if (o2hb_global_heartbeat_active()) { + reg->hr_region_num = find_first_zero_bit(o2hb_region_bitmap, + O2NM_MAX_REGIONS); + if (reg->hr_region_num >= O2NM_MAX_REGIONS) { + spin_unlock(&o2hb_live_lock); + return ERR_PTR(-EFBIG); + } + set_bit(reg->hr_region_num, o2hb_region_bitmap); + } list_add_tail(®->hr_all_item, &o2hb_all_regions); spin_unlock(&o2hb_live_lock); + config_item_init_type_name(®->hr_item, name, &o2hb_region_type); + + ret = o2hb_debug_region_init(reg, o2hb_debug_dir); + if (ret) { + config_item_put(®->hr_item); + return ERR_PTR(ret); + } + return ®->hr_item; } @@ -1612,6 +2000,10 @@ 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); + } hb_task = reg->hr_task; reg->hr_task = NULL; spin_unlock(&o2hb_live_lock); @@ -1628,6 +2020,9 @@ static void o2hb_heartbeat_group_drop_item(struct config_group *group, 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); } @@ -1688,6 +2083,41 @@ static ssize_t o2hb_heartbeat_group_threshold_store(struct o2hb_heartbeat_group return count; } +static +ssize_t o2hb_heartbeat_group_mode_show(struct o2hb_heartbeat_group *group, + char *page) +{ + return sprintf(page, "%s\n", + o2hb_heartbeat_mode_desc[o2hb_heartbeat_mode]); +} + +static +ssize_t o2hb_heartbeat_group_mode_store(struct o2hb_heartbeat_group *group, + const char *page, size_t count) +{ + unsigned int i; + int ret; + size_t len; + + len = (page[count - 1] == '\n') ? count - 1 : count; + if (!len) + return -EINVAL; + + for (i = 0; i < O2HB_HEARTBEAT_NUM_MODES; ++i) { + if (strnicmp(page, o2hb_heartbeat_mode_desc[i], len)) + continue; + + ret = o2hb_global_hearbeat_mode_set(i); + if (!ret) + printk(KERN_NOTICE "o2hb: Heartbeat mode set to %s\n", + o2hb_heartbeat_mode_desc[i]); + return count; + } + + return -EINVAL; + +} + static struct o2hb_heartbeat_group_attribute o2hb_heartbeat_group_attr_threshold = { .attr = { .ca_owner = THIS_MODULE, .ca_name = "dead_threshold", @@ -1696,8 +2126,17 @@ static struct o2hb_heartbeat_group_attribute o2hb_heartbeat_group_attr_threshold .store = o2hb_heartbeat_group_threshold_store, }; +static struct o2hb_heartbeat_group_attribute o2hb_heartbeat_group_attr_mode = { + .attr = { .ca_owner = THIS_MODULE, + .ca_name = "mode", + .ca_mode = S_IRUGO | S_IWUSR }, + .show = o2hb_heartbeat_group_mode_show, + .store = o2hb_heartbeat_group_mode_store, +}; + static struct configfs_attribute *o2hb_heartbeat_group_attrs[] = { &o2hb_heartbeat_group_attr_threshold.attr, + &o2hb_heartbeat_group_attr_mode.attr, NULL, }; @@ -1963,3 +2402,34 @@ void o2hb_stop_all_regions(void) spin_unlock(&o2hb_live_lock); } EXPORT_SYMBOL_GPL(o2hb_stop_all_regions); + +int o2hb_get_all_regions(char *region_uuids, u8 max_regions) +{ + struct o2hb_region *reg; + int numregs = 0; + char *p; + + spin_lock(&o2hb_live_lock); + + p = region_uuids; + list_for_each_entry(reg, &o2hb_all_regions, hr_all_item) { + mlog(0, "Region: %s\n", config_item_name(®->hr_item)); + if (numregs < max_regions) { + memcpy(p, config_item_name(®->hr_item), + O2HB_MAX_REGION_NAME_LEN); + p += O2HB_MAX_REGION_NAME_LEN; + } + numregs++; + } + + spin_unlock(&o2hb_live_lock); + + return numregs; +} +EXPORT_SYMBOL_GPL(o2hb_get_all_regions); + +int o2hb_global_heartbeat_active(void) +{ + return (o2hb_heartbeat_mode == O2HB_HEARTBEAT_GLOBAL); +} +EXPORT_SYMBOL(o2hb_global_heartbeat_active); diff --git a/fs/ocfs2/cluster/heartbeat.h b/fs/ocfs2/cluster/heartbeat.h index 2f1649253b4..00ad8e8fea5 100644 --- a/fs/ocfs2/cluster/heartbeat.h +++ b/fs/ocfs2/cluster/heartbeat.h @@ -31,6 +31,8 @@ #define O2HB_REGION_TIMEOUT_MS 2000 +#define O2HB_MAX_REGION_NAME_LEN 32 + /* number of changes to be seen as live */ #define O2HB_LIVE_THRESHOLD 2 /* number of equal samples to be seen as dead */ @@ -81,5 +83,7 @@ int o2hb_check_node_heartbeating(u8 node_num); int o2hb_check_node_heartbeating_from_callback(u8 node_num); int o2hb_check_local_node_heartbeating(void); void o2hb_stop_all_regions(void); +int o2hb_get_all_regions(char *region_uuids, u8 numregions); +int o2hb_global_heartbeat_active(void); #endif /* O2CLUSTER_HEARTBEAT_H */ diff --git a/fs/ocfs2/cluster/masklog.h b/fs/ocfs2/cluster/masklog.h index fd96e2a2fa5..ea2ed9f56c9 100644 --- a/fs/ocfs2/cluster/masklog.h +++ b/fs/ocfs2/cluster/masklog.h @@ -119,7 +119,8 @@ #define ML_ERROR 0x0000000100000000ULL /* sent to KERN_ERR */ #define ML_NOTICE 0x0000000200000000ULL /* setn to KERN_NOTICE */ #define ML_KTHREAD 0x0000000400000000ULL /* kernel thread activity */ -#define ML_RESERVATIONS 0x0000000800000000ULL /* ocfs2 alloc reservations */ +#define ML_RESERVATIONS 0x0000000800000000ULL /* ocfs2 alloc reservations */ +#define ML_CLUSTER 0x0000001000000000ULL /* cluster stack */ #define MLOG_INITIAL_AND_MASK (ML_ERROR|ML_NOTICE) #define MLOG_INITIAL_NOT_MASK (ML_ENTRY|ML_EXIT) diff --git a/fs/ocfs2/cluster/nodemanager.c b/fs/ocfs2/cluster/nodemanager.c index ed0c9f367fe..bb240647ca5 100644 --- a/fs/ocfs2/cluster/nodemanager.c +++ b/fs/ocfs2/cluster/nodemanager.c @@ -711,6 +711,8 @@ static struct config_item *o2nm_node_group_make_item(struct config_group *group, config_item_init_type_name(&node->nd_item, name, &o2nm_node_type); spin_lock_init(&node->nd_lock); + mlog(ML_CLUSTER, "o2nm: Registering node %s\n", name); + return &node->nd_item; } @@ -744,6 +746,9 @@ static void o2nm_node_group_drop_item(struct config_group *group, } write_unlock(&cluster->cl_nodes_lock); + mlog(ML_CLUSTER, "o2nm: Unregistered node %s\n", + config_item_name(&node->nd_item)); + config_item_put(item); } diff --git a/fs/ocfs2/cluster/ocfs2_nodemanager.h b/fs/ocfs2/cluster/ocfs2_nodemanager.h index 5b9854bad57..49b594325be 100644 --- a/fs/ocfs2/cluster/ocfs2_nodemanager.h +++ b/fs/ocfs2/cluster/ocfs2_nodemanager.h @@ -36,4 +36,10 @@ /* host name, group name, cluster name all 64 bytes */ #define O2NM_MAX_NAME_LEN 64 // __NEW_UTS_LEN +/* + * Maximum number of global heartbeat regions allowed. + * **CAUTION** Changing this number will break dlm compatibility. + */ +#define O2NM_MAX_REGIONS 32 + #endif /* _OCFS2_NODEMANAGER_H */ diff --git a/fs/ocfs2/cluster/tcp.c b/fs/ocfs2/cluster/tcp.c index aa75ca3f78d..9aa426e4212 100644 --- a/fs/ocfs2/cluster/tcp.c +++ b/fs/ocfs2/cluster/tcp.c @@ -977,7 +977,7 @@ static int o2net_tx_can_proceed(struct o2net_node *nn, int o2net_send_message_vec(u32 msg_type, u32 key, struct kvec *caller_vec, size_t caller_veclen, u8 target_node, int *status) { - int ret; + int ret = 0; struct o2net_msg *msg = NULL; size_t veclen, caller_bytes = 0; struct kvec *vec = NULL; @@ -1696,6 +1696,9 @@ static void o2net_hb_node_down_cb(struct o2nm_node *node, int node_num, { o2quo_hb_down(node_num); + if (!node) + return; + if (node_num != o2nm_this_node()) o2net_disconnect_node(node); @@ -1709,6 +1712,8 @@ static void o2net_hb_node_up_cb(struct o2nm_node *node, int node_num, o2quo_hb_up(node_num); + BUG_ON(!node); + /* ensure an immediate connect attempt */ nn->nn_last_connect_attempt = jiffies - (msecs_to_jiffies(o2net_reconnect_delay()) + 1); @@ -1759,6 +1764,7 @@ static int o2net_accept_one(struct socket *sock) struct sockaddr_in sin; struct socket *new_sock = NULL; struct o2nm_node *node = NULL; + struct o2nm_node *local_node = NULL; struct o2net_sock_container *sc = NULL; struct o2net_node *nn; @@ -1796,11 +1802,15 @@ static int o2net_accept_one(struct socket *sock) goto out; } - if (o2nm_this_node() > node->nd_num) { - mlog(ML_NOTICE, "unexpected connect attempted from a lower " - "numbered node '%s' at " "%pI4:%d with num %u\n", - node->nd_name, &sin.sin_addr.s_addr, - ntohs(sin.sin_port), node->nd_num); + 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)); ret = -EINVAL; goto out; } @@ -1857,6 +1867,8 @@ out: sock_release(new_sock); if (node) o2nm_node_put(node); + if (local_node) + o2nm_node_put(local_node); if (sc) sc_put(sc); return ret; diff --git a/fs/ocfs2/dcache.c b/fs/ocfs2/dcache.c index b4957c7d9fe..edaded48e7e 100644 --- a/fs/ocfs2/dcache.c +++ b/fs/ocfs2/dcache.c @@ -40,6 +40,14 @@ #include "inode.h" #include "super.h" +void ocfs2_dentry_attach_gen(struct dentry *dentry) +{ + unsigned long gen = + OCFS2_I(dentry->d_parent->d_inode)->ip_dir_lock_gen; + BUG_ON(dentry->d_inode); + dentry->d_fsdata = (void *)gen; +} + static int ocfs2_dentry_revalidate(struct dentry *dentry, struct nameidata *nd) @@ -51,11 +59,20 @@ static int ocfs2_dentry_revalidate(struct dentry *dentry, mlog_entry("(0x%p, '%.*s')\n", dentry, dentry->d_name.len, dentry->d_name.name); - /* Never trust a negative dentry - force a new lookup. */ + /* For a negative dentry - + * check the generation number of the parent and compare with the + * one stored in the inode. + */ if (inode == NULL) { - mlog(0, "negative dentry: %.*s\n", dentry->d_name.len, - dentry->d_name.name); - goto bail; + unsigned long gen = (unsigned long) dentry->d_fsdata; + unsigned long pgen = + OCFS2_I(dentry->d_parent->d_inode)->ip_dir_lock_gen; + mlog(0, "negative dentry: %.*s parent gen: %lu " + "dentry gen: %lu\n", + dentry->d_name.len, dentry->d_name.name, pgen, gen); + if (gen != pgen) + goto bail; + goto valid; } BUG_ON(!osb); @@ -96,6 +113,7 @@ static int ocfs2_dentry_revalidate(struct dentry *dentry, goto bail; } +valid: ret = 1; bail: @@ -227,6 +245,12 @@ int ocfs2_dentry_attach_lock(struct dentry *dentry, if (!inode) return 0; + if (!dentry->d_inode && dentry->d_fsdata) { + /* Converting a negative dentry to positive + Clear dentry->d_fsdata */ + dentry->d_fsdata = dl = NULL; + } + if (dl) { mlog_bug_on_msg(dl->dl_parent_blkno != parent_blkno, " \"%.*s\": old parent: %llu, new: %llu\n", @@ -452,6 +476,7 @@ static void ocfs2_dentry_iput(struct dentry *dentry, struct inode *inode) out: iput(inode); + ocfs2_dentry_attach_gen(dentry); } /* diff --git a/fs/ocfs2/dcache.h b/fs/ocfs2/dcache.h index f5dd1789acf..b79eff70995 100644 --- a/fs/ocfs2/dcache.h +++ b/fs/ocfs2/dcache.h @@ -64,5 +64,6 @@ void ocfs2_dentry_move(struct dentry *dentry, struct dentry *target, struct inode *old_dir, struct inode *new_dir); extern spinlock_t dentry_attach_lock; +void ocfs2_dentry_attach_gen(struct dentry *dentry); #endif /* OCFS2_DCACHE_H */ diff --git a/fs/ocfs2/dir.c b/fs/ocfs2/dir.c index f04ebcfffc4..c49f6de0e7a 100644 --- a/fs/ocfs2/dir.c +++ b/fs/ocfs2/dir.c @@ -3931,6 +3931,15 @@ static int ocfs2_dx_dir_rebalance(struct ocfs2_super *osb, struct inode *dir, goto out_commit; } + cpos = split_hash; + ret = ocfs2_dx_dir_new_cluster(dir, &et, cpos, handle, + data_ac, meta_ac, new_dx_leaves, + num_dx_leaves); + if (ret) { + mlog_errno(ret); + goto out_commit; + } + for (i = 0; i < num_dx_leaves; i++) { ret = ocfs2_journal_access_dl(handle, INODE_CACHE(dir), orig_dx_leaves[i], @@ -3939,15 +3948,14 @@ static int ocfs2_dx_dir_rebalance(struct ocfs2_super *osb, struct inode *dir, mlog_errno(ret); goto out_commit; } - } - cpos = split_hash; - ret = ocfs2_dx_dir_new_cluster(dir, &et, cpos, handle, - data_ac, meta_ac, new_dx_leaves, - num_dx_leaves); - if (ret) { - mlog_errno(ret); - goto out_commit; + ret = ocfs2_journal_access_dl(handle, INODE_CACHE(dir), + new_dx_leaves[i], + OCFS2_JOURNAL_ACCESS_WRITE); + if (ret) { + mlog_errno(ret); + goto out_commit; + } } ocfs2_dx_dir_transfer_leaf(dir, split_hash, handle, tmp_dx_leaf, diff --git a/fs/ocfs2/dlm/dlmcommon.h b/fs/ocfs2/dlm/dlmcommon.h index 4b6ae2c13b4..b36d0bf77a5 100644 --- a/fs/ocfs2/dlm/dlmcommon.h +++ b/fs/ocfs2/dlm/dlmcommon.h @@ -445,7 +445,9 @@ enum { DLM_LOCK_REQUEST_MSG, /* 515 */ DLM_RECO_DATA_DONE_MSG, /* 516 */ DLM_BEGIN_RECO_MSG, /* 517 */ - DLM_FINALIZE_RECO_MSG /* 518 */ + DLM_FINALIZE_RECO_MSG, /* 518 */ + DLM_QUERY_REGION, /* 519 */ + DLM_QUERY_NODEINFO, /* 520 */ }; struct dlm_reco_node_data @@ -727,6 +729,31 @@ struct dlm_cancel_join u8 domain[O2NM_MAX_NAME_LEN]; }; +struct dlm_query_region { + u8 qr_node; + u8 qr_numregions; + u8 qr_namelen; + u8 pad1; + u8 qr_domain[O2NM_MAX_NAME_LEN]; + u8 qr_regions[O2HB_MAX_REGION_NAME_LEN * O2NM_MAX_REGIONS]; +}; + +struct dlm_node_info { + u8 ni_nodenum; + u8 pad1; + u16 ni_ipv4_port; + u32 ni_ipv4_address; +}; + +struct dlm_query_nodeinfo { + u8 qn_nodenum; + u8 qn_numnodes; + u8 qn_namelen; + u8 pad1; + u8 qn_domain[O2NM_MAX_NAME_LEN]; + struct dlm_node_info qn_nodes[O2NM_MAX_NODES]; +}; + struct dlm_exit_domain { u8 node_idx; @@ -1030,6 +1057,7 @@ int dlm_drop_lockres_ref(struct dlm_ctxt *dlm, struct dlm_lock_resource *res); void dlm_clean_master_list(struct dlm_ctxt *dlm, u8 dead_node); +void dlm_force_free_mles(struct dlm_ctxt *dlm); int dlm_lock_basts_flushed(struct dlm_ctxt *dlm, struct dlm_lock *lock); int __dlm_lockres_has_locks(struct dlm_lock_resource *res); int __dlm_lockres_unused(struct dlm_lock_resource *res); diff --git a/fs/ocfs2/dlm/dlmdebug.c b/fs/ocfs2/dlm/dlmdebug.c index 5efdd37dfe4..272ec8631a5 100644 --- a/fs/ocfs2/dlm/dlmdebug.c +++ b/fs/ocfs2/dlm/dlmdebug.c @@ -493,7 +493,7 @@ static int debug_mle_print(struct dlm_ctxt *dlm, struct debug_buffer *db) struct hlist_head *bucket; struct hlist_node *list; int i, out = 0; - unsigned long total = 0, longest = 0, bktcnt; + unsigned long total = 0, longest = 0, bucket_count = 0; out += snprintf(db->buf + out, db->len - out, "Dumping MLEs for Domain: %s\n", dlm->name); @@ -505,13 +505,13 @@ static int debug_mle_print(struct dlm_ctxt *dlm, struct debug_buffer *db) mle = hlist_entry(list, struct dlm_master_list_entry, master_hash_node); ++total; - ++bktcnt; + ++bucket_count; if (db->len - out < 200) continue; out += dump_mle(mle, db->buf + out, db->len - out); } - longest = max(longest, bktcnt); - bktcnt = 0; + longest = max(longest, bucket_count); + bucket_count = 0; } spin_unlock(&dlm->master_lock); @@ -636,8 +636,14 @@ static void *lockres_seq_start(struct seq_file *m, loff_t *pos) spin_lock(&dlm->track_lock); if (oldres) track_list = &oldres->tracking; - else + else { track_list = &dlm->tracking_list; + if (list_empty(track_list)) { + dl = NULL; + spin_unlock(&dlm->track_lock); + goto bail; + } + } list_for_each_entry(res, track_list, tracking) { if (&res->tracking == &dlm->tracking_list) @@ -660,6 +666,7 @@ static void *lockres_seq_start(struct seq_file *m, loff_t *pos) } else dl = NULL; +bail: /* passed to seq_show */ return dl; } @@ -775,7 +782,9 @@ static int debug_state_print(struct dlm_ctxt *dlm, struct debug_buffer *db) /* Domain: xxxxxxxxxx Key: 0xdfbac769 */ out += snprintf(db->buf + out, db->len - out, - "Domain: %s Key: 0x%08x\n", dlm->name, dlm->key); + "Domain: %s Key: 0x%08x Protocol: %d.%d\n", + dlm->name, dlm->key, dlm->dlm_locking_proto.pv_major, + dlm->dlm_locking_proto.pv_minor); /* Thread Pid: xxx Node: xxx State: xxxxx */ out += snprintf(db->buf + out, db->len - out, diff --git a/fs/ocfs2/dlm/dlmdomain.c b/fs/ocfs2/dlm/dlmdomain.c index 153abb5abef..58a93b95373 100644 --- a/fs/ocfs2/dlm/dlmdomain.c +++ b/fs/ocfs2/dlm/dlmdomain.c @@ -128,10 +128,14 @@ static DECLARE_WAIT_QUEUE_HEAD(dlm_domain_events); * will have a negotiated version with the same major number and a minor * number equal or smaller. The dlm_ctxt->dlm_locking_proto field should * be used to determine what a running domain is actually using. + * + * New in version 1.1: + * - Message DLM_QUERY_REGION added to support global heartbeat + * - Message DLM_QUERY_NODEINFO added to allow online node removes */ static const struct dlm_protocol_version dlm_protocol = { .pv_major = 1, - .pv_minor = 0, + .pv_minor = 1, }; #define DLM_DOMAIN_BACKOFF_MS 200 @@ -142,6 +146,8 @@ static int dlm_assert_joined_handler(struct o2net_msg *msg, u32 len, void *data, void **ret_data); static int dlm_cancel_join_handler(struct o2net_msg *msg, u32 len, void *data, void **ret_data); +static int dlm_query_region_handler(struct o2net_msg *msg, u32 len, + void *data, void **ret_data); static int dlm_exit_domain_handler(struct o2net_msg *msg, u32 len, void *data, void **ret_data); static int dlm_protocol_compare(struct dlm_protocol_version *existing, @@ -693,6 +699,7 @@ void dlm_unregister_domain(struct dlm_ctxt *dlm) dlm_mark_domain_leaving(dlm); dlm_leave_domain(dlm); + dlm_force_free_mles(dlm); dlm_complete_dlm_shutdown(dlm); } dlm_put(dlm); @@ -920,6 +927,370 @@ static int dlm_assert_joined_handler(struct o2net_msg *msg, u32 len, void *data, return 0; } +static int dlm_match_regions(struct dlm_ctxt *dlm, + struct dlm_query_region *qr) +{ + char *local = NULL, *remote = qr->qr_regions; + char *l, *r; + int localnr, i, j, foundit; + int status = 0; + + if (!o2hb_global_heartbeat_active()) { + if (qr->qr_numregions) { + mlog(ML_ERROR, "Domain %s: Joining node %d has global " + "heartbeat enabled but local node %d does not\n", + qr->qr_domain, qr->qr_node, dlm->node_num); + status = -EINVAL; + } + goto bail; + } + + if (o2hb_global_heartbeat_active() && !qr->qr_numregions) { + mlog(ML_ERROR, "Domain %s: Local node %d has global " + "heartbeat enabled but joining node %d does not\n", + qr->qr_domain, dlm->node_num, qr->qr_node); + status = -EINVAL; + goto bail; + } + + r = remote; + for (i = 0; i < qr->qr_numregions; ++i) { + mlog(0, "Region %.*s\n", O2HB_MAX_REGION_NAME_LEN, r); + r += O2HB_MAX_REGION_NAME_LEN; + } + + local = kmalloc(sizeof(qr->qr_regions), GFP_KERNEL); + if (!local) { + status = -ENOMEM; + goto bail; + } + + localnr = o2hb_get_all_regions(local, O2NM_MAX_REGIONS); + + /* compare local regions with remote */ + l = local; + for (i = 0; i < localnr; ++i) { + foundit = 0; + r = remote; + for (j = 0; j <= qr->qr_numregions; ++j) { + if (!memcmp(l, r, O2HB_MAX_REGION_NAME_LEN)) { + foundit = 1; + break; + } + r += O2HB_MAX_REGION_NAME_LEN; + } + if (!foundit) { + status = -EINVAL; + mlog(ML_ERROR, "Domain %s: Region '%.*s' registered " + "in local node %d but not in joining node %d\n", + qr->qr_domain, O2HB_MAX_REGION_NAME_LEN, l, + dlm->node_num, qr->qr_node); + goto bail; + } + l += O2HB_MAX_REGION_NAME_LEN; + } + + /* compare remote with local regions */ + r = remote; + for (i = 0; i < qr->qr_numregions; ++i) { + foundit = 0; + l = local; + for (j = 0; j < localnr; ++j) { + if (!memcmp(r, l, O2HB_MAX_REGION_NAME_LEN)) { + foundit = 1; + break; + } + l += O2HB_MAX_REGION_NAME_LEN; + } + if (!foundit) { + status = -EINVAL; + mlog(ML_ERROR, "Domain %s: Region '%.*s' registered " + "in joining node %d but not in local node %d\n", + qr->qr_domain, O2HB_MAX_REGION_NAME_LEN, r, + qr->qr_node, dlm->node_num); + goto bail; + } + r += O2HB_MAX_REGION_NAME_LEN; + } + +bail: + kfree(local); + + return status; +} + +static int dlm_send_regions(struct dlm_ctxt *dlm, unsigned long *node_map) +{ + struct dlm_query_region *qr = NULL; + int status, ret = 0, i; + char *p; + + if (find_next_bit(node_map, O2NM_MAX_NODES, 0) >= O2NM_MAX_NODES) + goto bail; + + qr = kzalloc(sizeof(struct dlm_query_region), GFP_KERNEL); + if (!qr) { + ret = -ENOMEM; + mlog_errno(ret); + goto bail; + } + + qr->qr_node = dlm->node_num; + qr->qr_namelen = strlen(dlm->name); + memcpy(qr->qr_domain, dlm->name, qr->qr_namelen); + /* if local hb, the numregions will be zero */ + if (o2hb_global_heartbeat_active()) + qr->qr_numregions = o2hb_get_all_regions(qr->qr_regions, + O2NM_MAX_REGIONS); + + p = qr->qr_regions; + for (i = 0; i < qr->qr_numregions; ++i, p += O2HB_MAX_REGION_NAME_LEN) + mlog(0, "Region %.*s\n", O2HB_MAX_REGION_NAME_LEN, p); + + i = -1; + while ((i = find_next_bit(node_map, O2NM_MAX_NODES, + i + 1)) < O2NM_MAX_NODES) { + if (i == dlm->node_num) + continue; + + mlog(0, "Sending regions to node %d\n", i); + + ret = o2net_send_message(DLM_QUERY_REGION, DLM_MOD_KEY, qr, + sizeof(struct dlm_query_region), + i, &status); + if (ret >= 0) + ret = status; + if (ret) { + mlog(ML_ERROR, "Region mismatch %d, node %d\n", + ret, i); + break; + } + } + +bail: + kfree(qr); + return ret; +} + +static int dlm_query_region_handler(struct o2net_msg *msg, u32 len, + void *data, void **ret_data) +{ + struct dlm_query_region *qr; + struct dlm_ctxt *dlm = NULL; + int status = 0; + int locked = 0; + + qr = (struct dlm_query_region *) msg->buf; + + mlog(0, "Node %u queries hb regions on domain %s\n", qr->qr_node, + qr->qr_domain); + + status = -EINVAL; + + spin_lock(&dlm_domain_lock); + dlm = __dlm_lookup_domain_full(qr->qr_domain, qr->qr_namelen); + 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; + } + + 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; + } + + /* Support for global heartbeat was added in 1.1 */ + if (dlm->dlm_locking_proto.pv_major == 1 && + dlm->dlm_locking_proto.pv_minor == 0) { + mlog(ML_ERROR, "Node %d queried hb regions on domain %s " + "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; + } + + status = dlm_match_regions(dlm, qr); + +bail: + if (locked) + spin_unlock(&dlm->spinlock); + spin_unlock(&dlm_domain_lock); + + return status; +} + +static int dlm_match_nodes(struct dlm_ctxt *dlm, struct dlm_query_nodeinfo *qn) +{ + struct o2nm_node *local; + struct dlm_node_info *remote; + int i, j; + int status = 0; + + for (j = 0; j < qn->qn_numnodes; ++j) + mlog(0, "Node %3d, %pI4:%u\n", qn->qn_nodes[j].ni_nodenum, + &(qn->qn_nodes[j].ni_ipv4_address), + ntohs(qn->qn_nodes[j].ni_ipv4_port)); + + for (i = 0; i < O2NM_MAX_NODES && !status; ++i) { + local = o2nm_get_node_by_num(i); + remote = NULL; + for (j = 0; j < qn->qn_numnodes; ++j) { + if (qn->qn_nodes[j].ni_nodenum == i) { + remote = &(qn->qn_nodes[j]); + break; + } + } + + if (!local && !remote) + continue; + + if ((local && !remote) || (!local && remote)) + status = -EINVAL; + + if (!status && + ((remote->ni_nodenum != local->nd_num) || + (remote->ni_ipv4_port != local->nd_ipv4_port) || + (remote->ni_ipv4_address != local->nd_ipv4_address))) + status = -EINVAL; + + if (status) { + if (remote && !local) + mlog(ML_ERROR, "Domain %s: Node %d (%pI4:%u) " + "registered in joining node %d but not in " + "local node %d\n", qn->qn_domain, + remote->ni_nodenum, + &(remote->ni_ipv4_address), + ntohs(remote->ni_ipv4_port), + qn->qn_nodenum, dlm->node_num); + if (local && !remote) + mlog(ML_ERROR, "Domain %s: Node %d (%pI4:%u) " + "registered in local node %d but not in " + "joining node %d\n", qn->qn_domain, + local->nd_num, &(local->nd_ipv4_address), + ntohs(local->nd_ipv4_port), + dlm->node_num, qn->qn_nodenum); + BUG_ON((!local && !remote)); + } + + if (local) + o2nm_node_put(local); + } + + return status; +} + +static int dlm_send_nodeinfo(struct dlm_ctxt *dlm, unsigned long *node_map) +{ + struct dlm_query_nodeinfo *qn = NULL; + struct o2nm_node *node; + int ret = 0, status, count, i; + + if (find_next_bit(node_map, O2NM_MAX_NODES, 0) >= O2NM_MAX_NODES) + goto bail; + + qn = kzalloc(sizeof(struct dlm_query_nodeinfo), GFP_KERNEL); + if (!qn) { + ret = -ENOMEM; + mlog_errno(ret); + goto bail; + } + + for (i = 0, count = 0; i < O2NM_MAX_NODES; ++i) { + node = o2nm_get_node_by_num(i); + if (!node) + continue; + qn->qn_nodes[count].ni_nodenum = node->nd_num; + qn->qn_nodes[count].ni_ipv4_port = node->nd_ipv4_port; + qn->qn_nodes[count].ni_ipv4_address = node->nd_ipv4_address; + mlog(0, "Node %3d, %pI4:%u\n", node->nd_num, + &(node->nd_ipv4_address), ntohs(node->nd_ipv4_port)); + ++count; + o2nm_node_put(node); + } + + qn->qn_nodenum = dlm->node_num; + qn->qn_numnodes = count; + qn->qn_namelen = strlen(dlm->name); + memcpy(qn->qn_domain, dlm->name, qn->qn_namelen); + + i = -1; + while ((i = find_next_bit(node_map, O2NM_MAX_NODES, + i + 1)) < O2NM_MAX_NODES) { + if (i == dlm->node_num) + continue; + + mlog(0, "Sending nodeinfo to node %d\n", i); + + ret = o2net_send_message(DLM_QUERY_NODEINFO, DLM_MOD_KEY, + qn, sizeof(struct dlm_query_nodeinfo), + i, &status); + if (ret >= 0) + ret = status; + if (ret) { + mlog(ML_ERROR, "node mismatch %d, node %d\n", ret, i); + break; + } + } + +bail: + kfree(qn); + return ret; +} + +static int dlm_query_nodeinfo_handler(struct o2net_msg *msg, u32 len, + void *data, void **ret_data) +{ + struct dlm_query_nodeinfo *qn; + struct dlm_ctxt *dlm = NULL; + int locked = 0, status = -EINVAL; + + qn = (struct dlm_query_nodeinfo *) msg->buf; + + mlog(0, "Node %u queries nodes on domain %s\n", qn->qn_nodenum, + qn->qn_domain); + + spin_lock(&dlm_domain_lock); + dlm = __dlm_lookup_domain_full(qn->qn_domain, qn->qn_namelen); + if (!dlm) { + mlog(ML_ERROR, "Node %d queried nodes on domain %s before " + "join domain\n", qn->qn_nodenum, qn->qn_domain); + goto bail; + } + + spin_lock(&dlm->spinlock); + locked = 1; + if (dlm->joining_node != qn->qn_nodenum) { + mlog(ML_ERROR, "Node %d queried nodes on domain %s but " + "joining node is %d\n", qn->qn_nodenum, qn->qn_domain, + dlm->joining_node); + goto bail; + } + + /* Support for node query was added in 1.1 */ + if (dlm->dlm_locking_proto.pv_major == 1 && + dlm->dlm_locking_proto.pv_minor == 0) { + mlog(ML_ERROR, "Node %d queried nodes on domain %s " + "but active dlm protocol is %d.%d\n", qn->qn_nodenum, + qn->qn_domain, dlm->dlm_locking_proto.pv_major, + dlm->dlm_locking_proto.pv_minor); + goto bail; + } + + status = dlm_match_nodes(dlm, qn); + +bail: + if (locked) + spin_unlock(&dlm->spinlock); + spin_unlock(&dlm_domain_lock); + + return status; +} + static int dlm_cancel_join_handler(struct o2net_msg *msg, u32 len, void *data, void **ret_data) { @@ -1240,6 +1611,20 @@ static int dlm_try_to_join_domain(struct dlm_ctxt *dlm) set_bit(dlm->node_num, dlm->domain_map); 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) { + status = dlm_send_nodeinfo(dlm, ctxt->yes_resp_map); + if (status) { + mlog_errno(status); + goto bail; + } + status = dlm_send_regions(dlm, ctxt->yes_resp_map); + if (status) { + mlog_errno(status); + goto bail; + } + } + dlm_send_join_asserts(dlm, ctxt->yes_resp_map); /* Joined state *must* be set before the joining node @@ -1806,7 +2191,21 @@ static int dlm_register_net_handlers(void) sizeof(struct dlm_cancel_join), dlm_cancel_join_handler, NULL, NULL, &dlm_join_handlers); + if (status) + goto bail; + + status = o2net_register_handler(DLM_QUERY_REGION, DLM_MOD_KEY, + sizeof(struct dlm_query_region), + dlm_query_region_handler, + NULL, NULL, &dlm_join_handlers); + if (status) + goto bail; + + status = o2net_register_handler(DLM_QUERY_NODEINFO, DLM_MOD_KEY, + sizeof(struct dlm_query_nodeinfo), + dlm_query_nodeinfo_handler, + NULL, NULL, &dlm_join_handlers); bail: if (status < 0) dlm_unregister_net_handlers(); diff --git a/fs/ocfs2/dlm/dlmmaster.c b/fs/ocfs2/dlm/dlmmaster.c index 94b97fc6a88..f564b0e5f80 100644 --- a/fs/ocfs2/dlm/dlmmaster.c +++ b/fs/ocfs2/dlm/dlmmaster.c @@ -511,8 +511,6 @@ static void dlm_lockres_release(struct kref *kref) atomic_dec(&dlm->res_cur_count); - dlm_put(dlm); - if (!hlist_unhashed(&res->hash_node) || !list_empty(&res->granted) || !list_empty(&res->converting) || @@ -585,8 +583,6 @@ static void dlm_init_lockres(struct dlm_ctxt *dlm, res->migration_pending = 0; res->inflight_locks = 0; - /* put in dlm_lockres_release */ - dlm_grab(dlm); res->dlm = dlm; kref_init(&res->refs); @@ -3050,8 +3046,6 @@ int dlm_migrate_request_handler(struct o2net_msg *msg, u32 len, void *data, /* check for pre-existing lock */ spin_lock(&dlm->spinlock); res = __dlm_lookup_lockres(dlm, name, namelen, hash); - spin_lock(&dlm->master_lock); - if (res) { spin_lock(&res->spinlock); if (res->state & DLM_LOCK_RES_RECOVERING) { @@ -3069,14 +3063,15 @@ int dlm_migrate_request_handler(struct o2net_msg *msg, u32 len, void *data, spin_unlock(&res->spinlock); } + spin_lock(&dlm->master_lock); /* ignore status. only nonzero status would BUG. */ ret = dlm_add_migration_mle(dlm, res, mle, &oldmle, name, namelen, migrate->new_master, migrate->master); -unlock: spin_unlock(&dlm->master_lock); +unlock: spin_unlock(&dlm->spinlock); if (oldmle) { @@ -3438,3 +3433,43 @@ void dlm_lockres_release_ast(struct dlm_ctxt *dlm, wake_up(&res->wq); wake_up(&dlm->migration_wq); } + +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; + + /* + * We notified all other nodes that we are exiting the domain and + * marked the dlm state to DLM_CTXT_LEAVING. If any mles are still + * around we force free them and wake any processes that are waiting + * on the mles + */ + spin_lock(&dlm->spinlock); + spin_lock(&dlm->master_lock); + + BUG_ON(dlm->dlm_state != DLM_CTXT_LEAVING); + BUG_ON((find_next_bit(dlm->domain_map, O2NM_MAX_NODES, 0) < O2NM_MAX_NODES)); + + 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); + if (mle->type != DLM_MLE_BLOCK) { + mlog(ML_ERROR, "bad mle: %p\n", mle); + dlm_print_one_mle(mle); + } + atomic_set(&mle->woken, 1); + wake_up(&mle->wq); + + __dlm_unlink_mle(dlm, mle); + __dlm_mle_detach_hb_events(dlm, mle); + __dlm_put_mle(mle); + } + } + spin_unlock(&dlm->master_lock); + spin_unlock(&dlm->spinlock); +} diff --git a/fs/ocfs2/dlm/dlmrecovery.c b/fs/ocfs2/dlm/dlmrecovery.c index 9dfaac73b36..aaaffbcbe91 100644 --- a/fs/ocfs2/dlm/dlmrecovery.c +++ b/fs/ocfs2/dlm/dlmrecovery.c @@ -1997,6 +1997,8 @@ void dlm_move_lockres_to_recovery_list(struct dlm_ctxt *dlm, struct list_head *queue; struct dlm_lock *lock, *next; + assert_spin_locked(&dlm->spinlock); + assert_spin_locked(&res->spinlock); res->state |= DLM_LOCK_RES_RECOVERING; if (!list_empty(&res->recovering)) { mlog(0, @@ -2326,19 +2328,15 @@ static void dlm_do_local_recovery_cleanup(struct dlm_ctxt *dlm, u8 dead_node) /* zero the lvb if necessary */ dlm_revalidate_lvb(dlm, res, dead_node); if (res->owner == dead_node) { - if (res->state & DLM_LOCK_RES_DROPPING_REF) - mlog(0, "%s:%.*s: owned by " - "dead node %u, this node was " - "dropping its ref when it died. " - "continue, dropping the flag.\n", - dlm->name, res->lockname.len, - res->lockname.name, dead_node); - - /* the wake_up for this will happen when the - * RECOVERING flag is dropped later */ - res->state &= ~DLM_LOCK_RES_DROPPING_REF; + if (res->state & DLM_LOCK_RES_DROPPING_REF) { + mlog(ML_NOTICE, "Ignore %.*s for " + "recovery as it is being freed\n", + res->lockname.len, + res->lockname.name); + } else + dlm_move_lockres_to_recovery_list(dlm, + res); - dlm_move_lockres_to_recovery_list(dlm, res); } else if (res->owner == dlm->node_num) { dlm_free_dead_locks(dlm, res, dead_node); __dlm_lockres_calc_usage(dlm, res); diff --git a/fs/ocfs2/dlm/dlmthread.c b/fs/ocfs2/dlm/dlmthread.c index d4f73ca68fe..2211acf33d9 100644 --- a/fs/ocfs2/dlm/dlmthread.c +++ b/fs/ocfs2/dlm/dlmthread.c @@ -92,19 +92,27 @@ int __dlm_lockres_has_locks(struct dlm_lock_resource *res) * truly ready to be freed. */ int __dlm_lockres_unused(struct dlm_lock_resource *res) { - if (!__dlm_lockres_has_locks(res) && - (list_empty(&res->dirty) && !(res->state & DLM_LOCK_RES_DIRTY))) { - /* try not to scan the bitmap unless the first two - * conditions are already true */ - int bit = find_next_bit(res->refmap, O2NM_MAX_NODES, 0); - if (bit >= O2NM_MAX_NODES) { - /* since the bit for dlm->node_num is not - * set, inflight_locks better be zero */ - BUG_ON(res->inflight_locks != 0); - return 1; - } - } - return 0; + int bit; + + if (__dlm_lockres_has_locks(res)) + return 0; + + if (!list_empty(&res->dirty) || res->state & DLM_LOCK_RES_DIRTY) + return 0; + + if (res->state & DLM_LOCK_RES_RECOVERING) + return 0; + + 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; } @@ -152,45 +160,25 @@ void dlm_lockres_calc_usage(struct dlm_ctxt *dlm, spin_unlock(&dlm->spinlock); } -static int dlm_purge_lockres(struct dlm_ctxt *dlm, +static void dlm_purge_lockres(struct dlm_ctxt *dlm, struct dlm_lock_resource *res) { int master; int ret = 0; - spin_lock(&res->spinlock); - if (!__dlm_lockres_unused(res)) { - mlog(0, "%s:%.*s: tried to purge but not unused\n", - dlm->name, res->lockname.len, res->lockname.name); - __dlm_print_one_lock_resource(res); - spin_unlock(&res->spinlock); - BUG(); - } - - if (res->state & DLM_LOCK_RES_MIGRATING) { - mlog(0, "%s:%.*s: Delay dropref as this lockres is " - "being remastered\n", dlm->name, res->lockname.len, - res->lockname.name); - /* Re-add the lockres to the end of the purge list */ - if (!list_empty(&res->purge)) { - list_del_init(&res->purge); - list_add_tail(&res->purge, &dlm->purge_list); - } - spin_unlock(&res->spinlock); - return 0; - } + assert_spin_locked(&dlm->spinlock); + assert_spin_locked(&res->spinlock); master = (res->owner == dlm->node_num); - if (!master) - res->state |= DLM_LOCK_RES_DROPPING_REF; - spin_unlock(&res->spinlock); mlog(0, "purging lockres %.*s, master = %d\n", res->lockname.len, res->lockname.name, master); if (!master) { + res->state |= DLM_LOCK_RES_DROPPING_REF; /* drop spinlock... retake below */ + spin_unlock(&res->spinlock); spin_unlock(&dlm->spinlock); spin_lock(&res->spinlock); @@ -208,31 +196,35 @@ static int dlm_purge_lockres(struct dlm_ctxt *dlm, mlog(0, "%s:%.*s: dlm_deref_lockres returned %d\n", dlm->name, res->lockname.len, res->lockname.name, ret); spin_lock(&dlm->spinlock); + spin_lock(&res->spinlock); } - spin_lock(&res->spinlock); if (!list_empty(&res->purge)) { mlog(0, "removing lockres %.*s:%p from purgelist, " "master = %d\n", res->lockname.len, res->lockname.name, res, master); list_del_init(&res->purge); - spin_unlock(&res->spinlock); dlm_lockres_put(res); dlm->purge_count--; - } else - spin_unlock(&res->spinlock); + } + + if (!__dlm_lockres_unused(res)) { + mlog(ML_ERROR, "found lockres %s:%.*s: in use after deref\n", + dlm->name, res->lockname.len, res->lockname.name); + __dlm_print_one_lock_resource(res); + BUG(); + } __dlm_unhash_lockres(res); /* lockres is not in the hash now. drop the flag and wake up * any processes waiting in dlm_get_lock_resource. */ if (!master) { - spin_lock(&res->spinlock); res->state &= ~DLM_LOCK_RES_DROPPING_REF; spin_unlock(&res->spinlock); wake_up(&res->wq); - } - return 0; + } else + spin_unlock(&res->spinlock); } static void dlm_run_purge_list(struct dlm_ctxt *dlm, @@ -251,17 +243,7 @@ static void dlm_run_purge_list(struct dlm_ctxt *dlm, lockres = list_entry(dlm->purge_list.next, struct dlm_lock_resource, purge); - /* Status of the lockres *might* change so double - * check. If the lockres is unused, holding the dlm - * spinlock will prevent people from getting and more - * refs on it -- there's no need to keep the lockres - * spinlock. */ spin_lock(&lockres->spinlock); - unused = __dlm_lockres_unused(lockres); - spin_unlock(&lockres->spinlock); - - if (!unused) - continue; purge_jiffies = lockres->last_used + msecs_to_jiffies(DLM_PURGE_INTERVAL_MS); @@ -273,15 +255,29 @@ static void dlm_run_purge_list(struct dlm_ctxt *dlm, * in tail order, we can stop at the first * unpurgable resource -- anyone added after * him will have a greater last_used value */ + spin_unlock(&lockres->spinlock); break; } + /* Status of the lockres *might* change so double + * check. If the lockres is unused, holding the dlm + * spinlock will prevent people from getting and more + * refs on it. */ + unused = __dlm_lockres_unused(lockres); + if (!unused || + (lockres->state & DLM_LOCK_RES_MIGRATING)) { + mlog(0, "lockres %s:%.*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); + spin_unlock(&lockres->spinlock); + continue; + } + dlm_lockres_get(lockres); - /* This may drop and reacquire the dlm spinlock if it - * has to do migration. */ - if (dlm_purge_lockres(dlm, lockres)) - BUG(); + dlm_purge_lockres(dlm, lockres); dlm_lockres_put(lockres); diff --git a/fs/ocfs2/dlmfs/dlmfs.c b/fs/ocfs2/dlmfs/dlmfs.c index c2903b84bb7..a7ebd9d42dc 100644 --- a/fs/ocfs2/dlmfs/dlmfs.c +++ b/fs/ocfs2/dlmfs/dlmfs.c @@ -612,6 +612,7 @@ static const struct file_operations dlmfs_file_operations = { .poll = dlmfs_file_poll, .read = dlmfs_file_read, .write = dlmfs_file_write, + .llseek = default_llseek, }; static const struct inode_operations dlmfs_dir_inode_operations = { diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c index 5e02a893f46..e8d94d722ec 100644 --- a/fs/ocfs2/dlmglue.c +++ b/fs/ocfs2/dlmglue.c @@ -3635,10 +3635,18 @@ static int ocfs2_data_convert_worker(struct ocfs2_lock_res *lockres, { struct inode *inode; struct address_space *mapping; + struct ocfs2_inode_info *oi; inode = ocfs2_lock_res_inode(lockres); mapping = inode->i_mapping; + if (S_ISDIR(inode->i_mode)) { + oi = OCFS2_I(inode); + oi->ip_dir_lock_gen++; + mlog(0, "generation: %u\n", oi->ip_dir_lock_gen); + goto out; + } + if (!S_ISREG(inode->i_mode)) goto out; diff --git a/fs/ocfs2/dlmglue.h b/fs/ocfs2/dlmglue.h index d1ce48e1b3d..1d596d8c4a4 100644 --- a/fs/ocfs2/dlmglue.h +++ b/fs/ocfs2/dlmglue.h @@ -84,6 +84,7 @@ enum { OI_LS_PARENT, OI_LS_RENAME1, OI_LS_RENAME2, + OI_LS_REFLINK_TARGET, }; int ocfs2_dlm_init(struct ocfs2_super *osb); diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c index 81296b4e364..1ca6867935b 100644 --- a/fs/ocfs2/file.c +++ b/fs/ocfs2/file.c @@ -36,6 +36,7 @@ #include <linux/writeback.h> #include <linux/falloc.h> #include <linux/quotaops.h> +#include <linux/blkdev.h> #define MLOG_MASK_PREFIX ML_INODE #include <cluster/masklog.h> @@ -63,12 +64,6 @@ #include "buffer_head_io.h" -static int ocfs2_sync_inode(struct inode *inode) -{ - filemap_fdatawrite(inode->i_mapping); - return sync_mapping_buffers(inode->i_mapping); -} - static int ocfs2_init_file_private(struct inode *inode, struct file *file) { struct ocfs2_file_private *fp; @@ -179,19 +174,22 @@ static int ocfs2_sync_file(struct file *file, int datasync) { int err = 0; journal_t *journal; - struct dentry *dentry = file->f_path.dentry; struct inode *inode = file->f_mapping->host; struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); - mlog_entry("(0x%p, 0x%p, %d, '%.*s')\n", file, dentry, datasync, - dentry->d_name.len, dentry->d_name.name); - - err = ocfs2_sync_inode(dentry->d_inode); - if (err) - goto bail; + mlog_entry("(0x%p, %d, 0x%p, '%.*s')\n", file, datasync, + file->f_path.dentry, file->f_path.dentry->d_name.len, + file->f_path.dentry->d_name.name); - if (datasync && !(inode->i_state & I_DIRTY_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; + } journal = osb->journal->j_journal; err = jbd2_journal_force_commit(journal); @@ -361,7 +359,7 @@ static int ocfs2_cow_file_pos(struct inode *inode, if (!(ext_flags & OCFS2_EXT_REFCOUNTED)) goto out; - return ocfs2_refcount_cow(inode, fe_bh, cpos, 1, cpos+1); + return ocfs2_refcount_cow(inode, NULL, fe_bh, cpos, 1, cpos+1); out: return status; @@ -774,7 +772,7 @@ static int ocfs2_write_zero_page(struct inode *inode, u64 abs_from, BUG_ON(abs_to > (((u64)index + 1) << PAGE_CACHE_SHIFT)); BUG_ON(abs_from & (inode->i_blkbits - 1)); - page = grab_cache_page(mapping, index); + page = find_or_create_page(mapping, index, GFP_NOFS); if (!page) { ret = -ENOMEM; mlog_errno(ret); @@ -904,8 +902,8 @@ static int ocfs2_zero_extend_get_range(struct inode *inode, zero_clusters = last_cpos - zero_cpos; if (needs_cow) { - rc = ocfs2_refcount_cow(inode, di_bh, zero_cpos, zero_clusters, - UINT_MAX); + rc = ocfs2_refcount_cow(inode, NULL, di_bh, zero_cpos, + zero_clusters, UINT_MAX); if (rc) { mlog_errno(rc); goto out; @@ -2053,6 +2051,7 @@ out: } static int ocfs2_prepare_inode_for_refcount(struct inode *inode, + struct file *file, loff_t pos, size_t count, int *meta_level) { @@ -2070,7 +2069,7 @@ static int ocfs2_prepare_inode_for_refcount(struct inode *inode, *meta_level = 1; - ret = ocfs2_refcount_cow(inode, di_bh, cpos, clusters, UINT_MAX); + ret = ocfs2_refcount_cow(inode, file, di_bh, cpos, clusters, UINT_MAX); if (ret) mlog_errno(ret); out: @@ -2078,7 +2077,7 @@ out: return ret; } -static int ocfs2_prepare_inode_for_write(struct dentry *dentry, +static int ocfs2_prepare_inode_for_write(struct file *file, loff_t *ppos, size_t count, int appending, @@ -2086,6 +2085,7 @@ static int ocfs2_prepare_inode_for_write(struct dentry *dentry, int *has_refcount) { int ret = 0, meta_level = 0; + struct dentry *dentry = file->f_path.dentry; struct inode *inode = dentry->d_inode; loff_t saved_pos, end; @@ -2141,6 +2141,7 @@ static int ocfs2_prepare_inode_for_write(struct dentry *dentry, meta_level = -1; ret = ocfs2_prepare_inode_for_refcount(inode, + file, saved_pos, count, &meta_level); @@ -2223,6 +2224,8 @@ static ssize_t ocfs2_file_aio_write(struct kiocb *iocb, struct file *file = iocb->ki_filp; struct inode *inode = file->f_path.dentry->d_inode; struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); + int full_coherency = !(osb->s_mount_opt & + OCFS2_MOUNT_COHERENCY_BUFFERED); mlog_entry("(0x%p, %u, '%.*s')\n", file, (unsigned int)nr_segs, @@ -2246,16 +2249,39 @@ relock: have_alloc_sem = 1; } - /* concurrent O_DIRECT writes are allowed */ - rw_level = !direct_io; + /* + * Concurrent O_DIRECT writes are allowed with + * mount_option "coherency=buffered". + */ + rw_level = (!direct_io || full_coherency); + ret = ocfs2_rw_lock(inode, rw_level); if (ret < 0) { mlog_errno(ret); goto out_sems; } + /* + * O_DIRECT writes with "coherency=full" need to take EX cluster + * inode_lock to guarantee coherency. + */ + if (direct_io && full_coherency) { + /* + * We need to take and drop the inode lock to force + * other nodes to drop their caches. Buffered I/O + * already does this in write_begin(). + */ + ret = ocfs2_inode_lock(inode, NULL, 1); + if (ret < 0) { + mlog_errno(ret); + goto out_sems; + } + + ocfs2_inode_unlock(inode, 1); + } + can_do_direct = direct_io; - ret = ocfs2_prepare_inode_for_write(file->f_path.dentry, ppos, + ret = ocfs2_prepare_inode_for_write(file, ppos, iocb->ki_left, appending, &can_do_direct, &has_refcount); if (ret < 0) { @@ -2303,17 +2329,6 @@ relock: written = generic_file_direct_write(iocb, iov, &nr_segs, *ppos, ppos, count, ocount); if (written < 0) { - /* - * direct write may have instantiated a few - * blocks outside i_size. Trim these off again. - * Don't need i_size_read because we hold i_mutex. - * - * XXX(truncate): this looks buggy because ocfs2 did not - * actually implement ->truncate. Take a look at - * the new truncate sequence and update this accordingly - */ - if (*ppos + count > inode->i_size) - truncate_setsize(inode, inode->i_size); ret = written; goto out_dio; } @@ -2329,7 +2344,7 @@ out_dio: BUG_ON(ret == -EIOCBQUEUED && !(file->f_flags & O_DIRECT)); if (((file->f_flags & O_DSYNC) && !direct_io) || IS_SYNC(inode) || - ((file->f_flags & O_DIRECT) && has_refcount)) { + ((file->f_flags & O_DIRECT) && !direct_io)) { ret = filemap_fdatawrite_range(file->f_mapping, pos, pos + count - 1); if (ret < 0) @@ -2385,7 +2400,7 @@ static int ocfs2_splice_to_file(struct pipe_inode_info *pipe, { int ret; - ret = ocfs2_prepare_inode_for_write(out->f_path.dentry, &sd->pos, + ret = ocfs2_prepare_inode_for_write(out, &sd->pos, sd->total_len, 0, NULL, NULL); if (ret < 0) { mlog_errno(ret); diff --git a/fs/ocfs2/inode.c b/fs/ocfs2/inode.c index 0492464916b..f935fd6600d 100644 --- a/fs/ocfs2/inode.c +++ b/fs/ocfs2/inode.c @@ -335,6 +335,7 @@ void ocfs2_populate_inode(struct inode *inode, struct ocfs2_dinode *fe, else inode->i_fop = &ocfs2_dops_no_plocks; i_size_write(inode, le64_to_cpu(fe->i_size)); + OCFS2_I(inode)->ip_dir_lock_gen = 1; break; case S_IFLNK: if (ocfs2_inode_is_fast_symlink(inode)) @@ -488,7 +489,11 @@ static int ocfs2_read_locked_inode(struct inode *inode, OCFS2_BH_IGNORE_CACHE); } else { status = ocfs2_read_blocks_sync(osb, args->fi_blkno, 1, &bh); - if (!status) + /* + * If buffer is in jbd, then its checksum may not have been + * computed as yet. + */ + if (!status && !buffer_jbd(bh)) status = ocfs2_validate_inode_block(osb->sb, bh); } if (status < 0) { diff --git a/fs/ocfs2/inode.h b/fs/ocfs2/inode.h index 6de5a869db3..1c508b149b3 100644 --- a/fs/ocfs2/inode.h +++ b/fs/ocfs2/inode.h @@ -46,30 +46,28 @@ struct ocfs2_inode_info /* These fields are protected by ip_lock */ spinlock_t ip_lock; u32 ip_open_count; - u32 ip_clusters; struct list_head ip_io_markers; + u32 ip_clusters; + u16 ip_dyn_features; struct mutex ip_io_mutex; - u32 ip_flags; /* see below */ u32 ip_attr; /* inode attributes */ - u16 ip_dyn_features; /* protected by recovery_lock. */ struct inode *ip_next_orphan; - u32 ip_dir_start_lookup; - struct ocfs2_caching_info ip_metadata_cache; - struct ocfs2_extent_map ip_extent_map; - struct inode vfs_inode; struct jbd2_inode ip_jinode; + u32 ip_dir_start_lookup; + /* Only valid if the inode is the dir. */ u32 ip_last_used_slot; u64 ip_last_used_group; + u32 ip_dir_lock_gen; struct ocfs2_alloc_reservation ip_la_data_resv; }; diff --git a/fs/ocfs2/ioctl.c b/fs/ocfs2/ioctl.c index 7d9d9c132ce..7a486819615 100644 --- a/fs/ocfs2/ioctl.c +++ b/fs/ocfs2/ioctl.c @@ -26,6 +26,26 @@ #include <linux/ext2_fs.h> +#define o2info_from_user(a, b) \ + copy_from_user(&(a), (b), sizeof(a)) +#define o2info_to_user(a, b) \ + copy_to_user((typeof(a) __user *)b, &(a), sizeof(a)) + +/* + * This call is void because we are already reporting an error that may + * 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, + 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 int ocfs2_get_inode_attr(struct inode *inode, unsigned *flags) { int status; @@ -109,6 +129,328 @@ bail: return status; } +int ocfs2_info_handle_blocksize(struct inode *inode, + struct ocfs2_info_request __user *req) +{ + int status = -EFAULT; + struct ocfs2_info_blocksize oib; + + if (o2info_from_user(oib, req)) + goto bail; + + oib.ib_blocksize = inode->i_sb->s_blocksize; + oib.ib_req.ir_flags |= OCFS2_INFO_FL_FILLED; + + if (o2info_to_user(oib, req)) + goto bail; + + status = 0; +bail: + if (status) + o2info_set_request_error(oib, req); + + return status; +} + +int ocfs2_info_handle_clustersize(struct inode *inode, + struct ocfs2_info_request __user *req) +{ + int status = -EFAULT; + struct ocfs2_info_clustersize oic; + struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); + + if (o2info_from_user(oic, req)) + goto bail; + + oic.ic_clustersize = osb->s_clustersize; + oic.ic_req.ir_flags |= OCFS2_INFO_FL_FILLED; + + if (o2info_to_user(oic, req)) + goto bail; + + status = 0; +bail: + if (status) + o2info_set_request_error(oic, req); + + return status; +} + +int ocfs2_info_handle_maxslots(struct inode *inode, + struct ocfs2_info_request __user *req) +{ + int status = -EFAULT; + struct ocfs2_info_maxslots oim; + struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); + + if (o2info_from_user(oim, req)) + goto bail; + + oim.im_max_slots = osb->max_slots; + oim.im_req.ir_flags |= OCFS2_INFO_FL_FILLED; + + if (o2info_to_user(oim, req)) + goto bail; + + status = 0; +bail: + if (status) + o2info_set_request_error(oim, req); + + return status; +} + +int ocfs2_info_handle_label(struct inode *inode, + struct ocfs2_info_request __user *req) +{ + int status = -EFAULT; + struct ocfs2_info_label oil; + struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); + + if (o2info_from_user(oil, req)) + goto bail; + + memcpy(oil.il_label, osb->vol_label, OCFS2_MAX_VOL_LABEL_LEN); + oil.il_req.ir_flags |= OCFS2_INFO_FL_FILLED; + + if (o2info_to_user(oil, req)) + goto bail; + + status = 0; +bail: + if (status) + o2info_set_request_error(oil, req); + + return status; +} + +int ocfs2_info_handle_uuid(struct inode *inode, + struct ocfs2_info_request __user *req) +{ + int status = -EFAULT; + struct ocfs2_info_uuid oiu; + struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); + + if (o2info_from_user(oiu, req)) + goto bail; + + memcpy(oiu.iu_uuid_str, osb->uuid_str, OCFS2_TEXT_UUID_LEN + 1); + oiu.iu_req.ir_flags |= OCFS2_INFO_FL_FILLED; + + if (o2info_to_user(oiu, req)) + goto bail; + + status = 0; +bail: + if (status) + o2info_set_request_error(oiu, req); + + return status; +} + +int ocfs2_info_handle_fs_features(struct inode *inode, + struct ocfs2_info_request __user *req) +{ + int status = -EFAULT; + struct ocfs2_info_fs_features oif; + struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); + + if (o2info_from_user(oif, req)) + goto bail; + + oif.if_compat_features = osb->s_feature_compat; + oif.if_incompat_features = osb->s_feature_incompat; + oif.if_ro_compat_features = osb->s_feature_ro_compat; + oif.if_req.ir_flags |= OCFS2_INFO_FL_FILLED; + + if (o2info_to_user(oif, req)) + goto bail; + + status = 0; +bail: + if (status) + o2info_set_request_error(oif, req); + + return status; +} + +int ocfs2_info_handle_journal_size(struct inode *inode, + struct ocfs2_info_request __user *req) +{ + int status = -EFAULT; + struct ocfs2_info_journal_size oij; + struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); + + if (o2info_from_user(oij, req)) + goto bail; + + oij.ij_journal_size = osb->journal->j_inode->i_size; + + oij.ij_req.ir_flags |= OCFS2_INFO_FL_FILLED; + + if (o2info_to_user(oij, req)) + goto bail; + + status = 0; +bail: + if (status) + o2info_set_request_error(oij, req); + + return status; +} + +int ocfs2_info_handle_unknown(struct inode *inode, + struct ocfs2_info_request __user *req) +{ + int status = -EFAULT; + struct ocfs2_info_request oir; + + if (o2info_from_user(oir, req)) + goto bail; + + oir.ir_flags &= ~OCFS2_INFO_FL_FILLED; + + if (o2info_to_user(oir, req)) + goto bail; + + status = 0; +bail: + if (status) + o2info_set_request_error(oir, req); + + return status; +} + +/* + * Validate and distinguish OCFS2_IOC_INFO requests. + * + * - validate the magic number. + * - distinguish different requests. + * - validate size of different requests. + */ +int ocfs2_info_handle_request(struct inode *inode, + struct ocfs2_info_request __user *req) +{ + int status = -EFAULT; + struct ocfs2_info_request oir; + + if (o2info_from_user(oir, req)) + goto bail; + + status = -EINVAL; + if (oir.ir_magic != OCFS2_INFO_MAGIC) + goto bail; + + switch (oir.ir_code) { + case OCFS2_INFO_BLOCKSIZE: + if (oir.ir_size == sizeof(struct ocfs2_info_blocksize)) + status = ocfs2_info_handle_blocksize(inode, req); + break; + case OCFS2_INFO_CLUSTERSIZE: + if (oir.ir_size == sizeof(struct ocfs2_info_clustersize)) + status = ocfs2_info_handle_clustersize(inode, req); + break; + case OCFS2_INFO_MAXSLOTS: + if (oir.ir_size == sizeof(struct ocfs2_info_maxslots)) + status = ocfs2_info_handle_maxslots(inode, req); + break; + case OCFS2_INFO_LABEL: + if (oir.ir_size == sizeof(struct ocfs2_info_label)) + status = ocfs2_info_handle_label(inode, req); + break; + case OCFS2_INFO_UUID: + if (oir.ir_size == sizeof(struct ocfs2_info_uuid)) + status = ocfs2_info_handle_uuid(inode, req); + break; + case OCFS2_INFO_FS_FEATURES: + if (oir.ir_size == sizeof(struct ocfs2_info_fs_features)) + status = ocfs2_info_handle_fs_features(inode, req); + break; + case OCFS2_INFO_JOURNAL_SIZE: + if (oir.ir_size == sizeof(struct ocfs2_info_journal_size)) + status = ocfs2_info_handle_journal_size(inode, req); + break; + default: + status = ocfs2_info_handle_unknown(inode, req); + break; + } + +bail: + return status; +} + +int ocfs2_get_request_ptr(struct ocfs2_info *info, int idx, + u64 *req_addr, int compat_flag) +{ + int status = -EFAULT; + u64 __user *bp = NULL; + + if (compat_flag) { +#ifdef CONFIG_COMPAT + /* + * pointer bp stores the base address of a pointers array, + * which collects all addresses of separate request. + */ + bp = (u64 __user *)(unsigned long)compat_ptr(info->oi_requests); +#else + BUG(); +#endif + } else + bp = (u64 __user *)(unsigned long)(info->oi_requests); + + if (o2info_from_user(*req_addr, bp + idx)) + goto bail; + + status = 0; +bail: + return status; +} + +/* + * OCFS2_IOC_INFO handles an array of requests passed from userspace. + * + * ocfs2_info_handle() recevies a large info aggregation, grab and + * validate the request count from header, then break it into small + * pieces, later specific handlers can handle them one by one. + * + * Idea here is to make each separate request small enough to ensure + * 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) +{ + int i, status = 0; + u64 req_addr; + struct ocfs2_info_request __user *reqp; + + if ((info->oi_count > OCFS2_INFO_MAX_REQUEST) || + (!info->oi_requests)) { + status = -EINVAL; + goto bail; + } + + for (i = 0; i < info->oi_count; i++) { + + status = ocfs2_get_request_ptr(info, i, &req_addr, compat_flag); + if (status) + break; + + reqp = (struct ocfs2_info_request *)(unsigned long)req_addr; + if (!reqp) { + status = -EINVAL; + goto bail; + } + + status = ocfs2_info_handle_request(inode, reqp); + if (status) + break; + } + +bail: + return status; +} + long ocfs2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) { struct inode *inode = filp->f_path.dentry->d_inode; @@ -120,6 +462,7 @@ long ocfs2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) struct reflink_arguments args; const char *old_path, *new_path; bool preserve; + struct ocfs2_info info; switch (cmd) { case OCFS2_IOC_GETFLAGS: @@ -174,6 +517,12 @@ long ocfs2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) 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))) + return -EFAULT; + + return ocfs2_info_handle(inode, &info, 0); default: return -ENOTTY; } @@ -185,6 +534,7 @@ 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 ocfs2_info info; switch (cmd) { case OCFS2_IOC32_GETFLAGS: @@ -209,6 +559,12 @@ long ocfs2_compat_ioctl(struct file *file, unsigned cmd, unsigned long arg) 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))) + return -EFAULT; + + return ocfs2_info_handle(inode, &info, 1); default: return -ENOIOCTLCMD; } diff --git a/fs/ocfs2/journal.c b/fs/ocfs2/journal.c index 9b57c0350ff..faa2303dbf0 100644 --- a/fs/ocfs2/journal.c +++ b/fs/ocfs2/journal.c @@ -301,7 +301,6 @@ static int ocfs2_commit_cache(struct ocfs2_super *osb) { int status = 0; unsigned int flushed; - unsigned long old_id; struct ocfs2_journal *journal = NULL; mlog_entry_void(); @@ -326,7 +325,7 @@ static int ocfs2_commit_cache(struct ocfs2_super *osb) goto finally; } - old_id = ocfs2_inc_trans_id(journal); + ocfs2_inc_trans_id(journal); flushed = atomic_read(&journal->j_num_trans); atomic_set(&journal->j_num_trans, 0); @@ -342,9 +341,6 @@ finally: return status; } -/* pass it NULL and it will allocate a new handle object for you. If - * you pass it a handle however, it may still return error, in which - * case it has free'd the passed handle for you. */ handle_t *ocfs2_start_trans(struct ocfs2_super *osb, int max_buffs) { journal_t *journal = osb->journal->j_journal; @@ -1888,6 +1884,8 @@ void ocfs2_queue_orphan_scan(struct ocfs2_super *osb) os = &osb->osb_orphan_scan; + mlog(0, "Begin orphan scan\n"); + if (atomic_read(&os->os_state) == ORPHAN_SCAN_INACTIVE) goto out; @@ -1920,6 +1918,7 @@ void ocfs2_queue_orphan_scan(struct ocfs2_super *osb) unlock: ocfs2_orphan_scan_unlock(osb, seqno); out: + mlog(0, "Orphan scan completed\n"); return; } diff --git a/fs/ocfs2/journal.h b/fs/ocfs2/journal.h index b5baaa8e710..43e56b97f9c 100644 --- a/fs/ocfs2/journal.h +++ b/fs/ocfs2/journal.h @@ -67,11 +67,12 @@ struct ocfs2_journal { struct buffer_head *j_bh; /* Journal disk inode block */ atomic_t j_num_trans; /* Number of transactions * currently in the system. */ + spinlock_t j_lock; unsigned long j_trans_id; struct rw_semaphore j_trans_barrier; wait_queue_head_t j_checkpointed; - spinlock_t j_lock; + /* both fields protected by j_lock*/ struct list_head j_la_cleanups; struct work_struct j_recovery_work; }; diff --git a/fs/ocfs2/mmap.c b/fs/ocfs2/mmap.c index af2b8fe1f13..7e32db9c2c9 100644 --- a/fs/ocfs2/mmap.c +++ b/fs/ocfs2/mmap.c @@ -59,10 +59,11 @@ static int ocfs2_fault(struct vm_area_struct *area, struct vm_fault *vmf) return ret; } -static int __ocfs2_page_mkwrite(struct inode *inode, struct buffer_head *di_bh, +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; struct address_space *mapping = inode->i_mapping; loff_t pos = page_offset(page); unsigned int len = PAGE_CACHE_SIZE; @@ -74,9 +75,11 @@ static int __ocfs2_page_mkwrite(struct inode *inode, struct buffer_head *di_bh, /* * 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 >> PAGE_CACHE_SHIFT; - if (page->index > last_index) { + last_index = (size - 1) >> PAGE_CACHE_SHIFT; + if (unlikely(!size || page->index > last_index)) { ret = -EINVAL; goto out; } @@ -107,9 +110,9 @@ static int __ocfs2_page_mkwrite(struct inode *inode, struct buffer_head *di_bh, * because the "write" would invalidate their data. */ if (page->index == last_index) - len = size & ~PAGE_CACHE_MASK; + len = ((size - 1) & ~PAGE_CACHE_MASK) + 1; - ret = ocfs2_write_begin_nolock(mapping, pos, len, 0, &locked_page, + ret = ocfs2_write_begin_nolock(file, mapping, pos, len, 0, &locked_page, &fsdata, di_bh, page); if (ret) { if (ret != -ENOSPC) @@ -157,7 +160,7 @@ static int ocfs2_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) */ down_write(&OCFS2_I(inode)->ip_alloc_sem); - ret = __ocfs2_page_mkwrite(inode, di_bh, page); + ret = __ocfs2_page_mkwrite(vma->vm_file, di_bh, page); up_write(&OCFS2_I(inode)->ip_alloc_sem); diff --git a/fs/ocfs2/namei.c b/fs/ocfs2/namei.c index f171b51a74f..e7bde21149a 100644 --- a/fs/ocfs2/namei.c +++ b/fs/ocfs2/namei.c @@ -171,7 +171,8 @@ bail_add: ret = ERR_PTR(status); goto bail_unlock; } - } + } else + ocfs2_dentry_attach_gen(dentry); bail_unlock: /* Don't drop the cluster lock until *after* the d_add -- @@ -472,32 +473,23 @@ leave: return status; } -static int ocfs2_mknod_locked(struct ocfs2_super *osb, - struct inode *dir, - struct inode *inode, - dev_t dev, - struct buffer_head **new_fe_bh, - struct buffer_head *parent_fe_bh, - handle_t *handle, - struct ocfs2_alloc_context *inode_ac) +static int __ocfs2_mknod_locked(struct inode *dir, + struct inode *inode, + dev_t dev, + struct buffer_head **new_fe_bh, + struct buffer_head *parent_fe_bh, + handle_t *handle, + struct ocfs2_alloc_context *inode_ac, + u64 fe_blkno, u64 suballoc_loc, u16 suballoc_bit) { int status = 0; + struct ocfs2_super *osb = OCFS2_SB(dir->i_sb); struct ocfs2_dinode *fe = NULL; struct ocfs2_extent_list *fel; - u64 suballoc_loc, fe_blkno = 0; - u16 suballoc_bit; u16 feat; *new_fe_bh = NULL; - status = ocfs2_claim_new_inode(handle, dir, parent_fe_bh, - inode_ac, &suballoc_loc, - &suballoc_bit, &fe_blkno); - if (status < 0) { - mlog_errno(status); - goto leave; - } - /* populate as many fields early on as possible - many of * these are used by the support functions here and in * callers. */ @@ -591,6 +583,34 @@ leave: return status; } +static int ocfs2_mknod_locked(struct ocfs2_super *osb, + struct inode *dir, + struct inode *inode, + dev_t dev, + struct buffer_head **new_fe_bh, + struct buffer_head *parent_fe_bh, + handle_t *handle, + struct ocfs2_alloc_context *inode_ac) +{ + int status = 0; + u64 suballoc_loc, fe_blkno = 0; + u16 suballoc_bit; + + *new_fe_bh = NULL; + + status = ocfs2_claim_new_inode(handle, dir, parent_fe_bh, + inode_ac, &suballoc_loc, + &suballoc_bit, &fe_blkno); + if (status < 0) { + mlog_errno(status); + return status; + } + + return __ocfs2_mknod_locked(dir, inode, dev, new_fe_bh, + parent_fe_bh, handle, inode_ac, + fe_blkno, suballoc_loc, suballoc_bit); +} + static int ocfs2_mkdir(struct inode *dir, struct dentry *dentry, int mode) @@ -1852,61 +1872,117 @@ bail: return status; } -static int ocfs2_prepare_orphan_dir(struct ocfs2_super *osb, - struct inode **ret_orphan_dir, - u64 blkno, - char *name, - struct ocfs2_dir_lookup_result *lookup) +static int ocfs2_lookup_lock_orphan_dir(struct ocfs2_super *osb, + struct inode **ret_orphan_dir, + struct buffer_head **ret_orphan_dir_bh) { struct inode *orphan_dir_inode; struct buffer_head *orphan_dir_bh = NULL; - int status = 0; - - status = ocfs2_blkno_stringify(blkno, name); - if (status < 0) { - mlog_errno(status); - return status; - } + int ret = 0; orphan_dir_inode = ocfs2_get_system_file_inode(osb, ORPHAN_DIR_SYSTEM_INODE, osb->slot_num); if (!orphan_dir_inode) { - status = -ENOENT; - mlog_errno(status); - return status; + ret = -ENOENT; + mlog_errno(ret); + return ret; } mutex_lock(&orphan_dir_inode->i_mutex); - status = ocfs2_inode_lock(orphan_dir_inode, &orphan_dir_bh, 1); - if (status < 0) { - mlog_errno(status); - goto leave; + ret = ocfs2_inode_lock(orphan_dir_inode, &orphan_dir_bh, 1); + if (ret < 0) { + mutex_unlock(&orphan_dir_inode->i_mutex); + iput(orphan_dir_inode); + + mlog_errno(ret); + return ret; } - status = ocfs2_prepare_dir_for_insert(osb, orphan_dir_inode, - orphan_dir_bh, name, - OCFS2_ORPHAN_NAMELEN, lookup); - if (status < 0) { - ocfs2_inode_unlock(orphan_dir_inode, 1); + *ret_orphan_dir = orphan_dir_inode; + *ret_orphan_dir_bh = orphan_dir_bh; - mlog_errno(status); - goto leave; + return 0; +} + +static int __ocfs2_prepare_orphan_dir(struct inode *orphan_dir_inode, + struct buffer_head *orphan_dir_bh, + u64 blkno, + char *name, + struct ocfs2_dir_lookup_result *lookup) +{ + int ret; + struct ocfs2_super *osb = OCFS2_SB(orphan_dir_inode->i_sb); + + ret = ocfs2_blkno_stringify(blkno, name); + if (ret < 0) { + mlog_errno(ret); + return ret; + } + + ret = ocfs2_prepare_dir_for_insert(osb, orphan_dir_inode, + orphan_dir_bh, name, + OCFS2_ORPHAN_NAMELEN, lookup); + if (ret < 0) { + mlog_errno(ret); + return ret; + } + + return 0; +} + +/** + * ocfs2_prepare_orphan_dir() - Prepare an orphan directory for + * insertion of an orphan. + * @osb: ocfs2 file system + * @ret_orphan_dir: Orphan dir inode - returned locked! + * @blkno: Actual block number of the inode to be inserted into orphan dir. + * @lookup: dir lookup result, to be passed back into functions like + * ocfs2_orphan_add + * + * Returns zero on success and the ret_orphan_dir, name and lookup + * fields will be populated. + * + * Returns non-zero on failure. + */ +static int ocfs2_prepare_orphan_dir(struct ocfs2_super *osb, + struct inode **ret_orphan_dir, + u64 blkno, + char *name, + struct ocfs2_dir_lookup_result *lookup) +{ + struct inode *orphan_dir_inode = NULL; + struct buffer_head *orphan_dir_bh = NULL; + int ret = 0; + + ret = ocfs2_lookup_lock_orphan_dir(osb, &orphan_dir_inode, + &orphan_dir_bh); + if (ret < 0) { + mlog_errno(ret); + return ret; + } + + ret = __ocfs2_prepare_orphan_dir(orphan_dir_inode, orphan_dir_bh, + blkno, name, lookup); + if (ret < 0) { + mlog_errno(ret); + goto out; } *ret_orphan_dir = orphan_dir_inode; -leave: - if (status) { +out: + brelse(orphan_dir_bh); + + if (ret) { + ocfs2_inode_unlock(orphan_dir_inode, 1); mutex_unlock(&orphan_dir_inode->i_mutex); iput(orphan_dir_inode); } - brelse(orphan_dir_bh); - - mlog_exit(status); - return status; + mlog_exit(ret); + return ret; } static int ocfs2_orphan_add(struct ocfs2_super *osb, @@ -2053,6 +2129,99 @@ leave: return status; } +/** + * ocfs2_prep_new_orphaned_file() - Prepare the orphan dir to recieve a newly + * allocated file. This is different from the typical 'add to orphan dir' + * operation in that the inode does not yet exist. This is a problem because + * the orphan dir stringifies the inode block number to come up with it's + * dirent. Obviously if the inode does not yet exist we have a chicken and egg + * problem. This function works around it by calling deeper into the orphan + * and suballoc code than other callers. Use this only by necessity. + * @dir: The directory which this inode will ultimately wind up under - not the + * orphan dir! + * @dir_bh: buffer_head the @dir inode block + * @orphan_name: string of length (CFS2_ORPHAN_NAMELEN + 1). Will be filled + * with the string to be used for orphan dirent. Pass back to the orphan dir + * code. + * @ret_orphan_dir: orphan dir inode returned to be passed back into orphan + * dir code. + * @ret_di_blkno: block number where the new inode will be allocated. + * @orphan_insert: Dir insert context to be passed back into orphan dir code. + * @ret_inode_ac: Inode alloc context to be passed back to the allocator. + * + * Returns zero on success and the ret_orphan_dir, name and lookup + * fields will be populated. + * + * Returns non-zero on failure. + */ +static int ocfs2_prep_new_orphaned_file(struct inode *dir, + struct buffer_head *dir_bh, + char *orphan_name, + struct inode **ret_orphan_dir, + u64 *ret_di_blkno, + struct ocfs2_dir_lookup_result *orphan_insert, + struct ocfs2_alloc_context **ret_inode_ac) +{ + int ret; + u64 di_blkno; + struct ocfs2_super *osb = OCFS2_SB(dir->i_sb); + struct inode *orphan_dir = NULL; + struct buffer_head *orphan_dir_bh = NULL; + struct ocfs2_alloc_context *inode_ac = NULL; + + ret = ocfs2_lookup_lock_orphan_dir(osb, &orphan_dir, &orphan_dir_bh); + if (ret < 0) { + mlog_errno(ret); + return ret; + } + + /* reserve an inode spot */ + ret = ocfs2_reserve_new_inode(osb, &inode_ac); + if (ret < 0) { + if (ret != -ENOSPC) + mlog_errno(ret); + goto out; + } + + ret = ocfs2_find_new_inode_loc(dir, dir_bh, inode_ac, + &di_blkno); + if (ret) { + mlog_errno(ret); + goto out; + } + + ret = __ocfs2_prepare_orphan_dir(orphan_dir, orphan_dir_bh, + di_blkno, orphan_name, orphan_insert); + if (ret < 0) { + mlog_errno(ret); + goto out; + } + +out: + if (ret == 0) { + *ret_orphan_dir = orphan_dir; + *ret_di_blkno = di_blkno; + *ret_inode_ac = inode_ac; + /* + * orphan_name and orphan_insert are already up to + * date via prepare_orphan_dir + */ + } else { + /* Unroll reserve_new_inode* */ + if (inode_ac) + ocfs2_free_alloc_context(inode_ac); + + /* Unroll orphan dir locking */ + mutex_unlock(&orphan_dir->i_mutex); + ocfs2_inode_unlock(orphan_dir, 1); + iput(orphan_dir); + } + + brelse(orphan_dir_bh); + + return 0; +} + int ocfs2_create_inode_in_orphan(struct inode *dir, int mode, struct inode **new_inode) @@ -2068,6 +2237,8 @@ int ocfs2_create_inode_in_orphan(struct inode *dir, struct buffer_head *new_di_bh = NULL; struct ocfs2_alloc_context *inode_ac = NULL; struct ocfs2_dir_lookup_result orphan_insert = { NULL, }; + u64 uninitialized_var(di_blkno), suballoc_loc; + u16 suballoc_bit; status = ocfs2_inode_lock(dir, &parent_di_bh, 1); if (status < 0) { @@ -2076,20 +2247,9 @@ int ocfs2_create_inode_in_orphan(struct inode *dir, return status; } - /* - * We give the orphan dir the root blkno to fake an orphan name, - * and allocate enough space for our insertion. - */ - status = ocfs2_prepare_orphan_dir(osb, &orphan_dir, - osb->root_blkno, - orphan_name, &orphan_insert); - if (status < 0) { - mlog_errno(status); - goto leave; - } - - /* reserve an inode spot */ - status = ocfs2_reserve_new_inode(osb, &inode_ac); + status = ocfs2_prep_new_orphaned_file(dir, parent_di_bh, + orphan_name, &orphan_dir, + &di_blkno, &orphan_insert, &inode_ac); if (status < 0) { if (status != -ENOSPC) mlog_errno(status); @@ -2116,17 +2276,20 @@ int ocfs2_create_inode_in_orphan(struct inode *dir, goto leave; did_quota_inode = 1; - inode->i_nlink = 0; - /* do the real work now. */ - status = ocfs2_mknod_locked(osb, dir, inode, - 0, &new_di_bh, parent_di_bh, handle, - inode_ac); + status = ocfs2_claim_new_inode_at_loc(handle, dir, inode_ac, + &suballoc_loc, + &suballoc_bit, di_blkno); if (status < 0) { mlog_errno(status); goto leave; } - status = ocfs2_blkno_stringify(OCFS2_I(inode)->ip_blkno, orphan_name); + inode->i_nlink = 0; + /* do the real work now. */ + status = __ocfs2_mknod_locked(dir, inode, + 0, &new_di_bh, parent_di_bh, handle, + inode_ac, di_blkno, suballoc_loc, + suballoc_bit); if (status < 0) { mlog_errno(status); goto leave; diff --git a/fs/ocfs2/ocfs2.h b/fs/ocfs2/ocfs2.h index c67003b6b5a..d8408217e3b 100644 --- a/fs/ocfs2/ocfs2.h +++ b/fs/ocfs2/ocfs2.h @@ -150,26 +150,33 @@ typedef void (*ocfs2_lock_callback)(int status, unsigned long data); struct ocfs2_lock_res { void *l_priv; struct ocfs2_lock_res_ops *l_ops; - spinlock_t l_lock; + struct list_head l_blocked_list; struct list_head l_mask_waiters; - enum ocfs2_lock_type l_type; unsigned long l_flags; char l_name[OCFS2_LOCK_ID_MAX_LEN]; - int l_level; unsigned int l_ro_holders; unsigned int l_ex_holders; - struct ocfs2_dlm_lksb l_lksb; + unsigned char l_level; + + /* Data packed - type enum ocfs2_lock_type */ + unsigned char l_type; /* used from AST/BAST funcs. */ - enum ocfs2_ast_action l_action; - enum ocfs2_unlock_action l_unlock_action; - int l_requested; - int l_blocking; + /* Data packed - enum type ocfs2_ast_action */ + unsigned char l_action; + /* Data packed - enum type ocfs2_unlock_action */ + unsigned char l_unlock_action; + unsigned char l_requested; + unsigned char l_blocking; unsigned int l_pending_gen; + spinlock_t l_lock; + + struct ocfs2_dlm_lksb l_lksb; + wait_queue_head_t l_event; struct list_head l_debug_list; @@ -243,7 +250,7 @@ enum ocfs2_local_alloc_state enum ocfs2_mount_options { - OCFS2_MOUNT_HB_LOCAL = 1 << 0, /* Heartbeat started in local mode */ + OCFS2_MOUNT_HB_LOCAL = 1 << 0, /* Local heartbeat */ OCFS2_MOUNT_BARRIER = 1 << 1, /* Use block barriers */ OCFS2_MOUNT_NOINTR = 1 << 2, /* Don't catch signals */ OCFS2_MOUNT_ERRORS_PANIC = 1 << 3, /* Panic on errors */ @@ -256,6 +263,10 @@ enum ocfs2_mount_options control lists */ OCFS2_MOUNT_USRQUOTA = 1 << 10, /* We support user quotas */ OCFS2_MOUNT_GRPQUOTA = 1 << 11, /* We support group quotas */ + OCFS2_MOUNT_COHERENCY_BUFFERED = 1 << 12, /* Allow concurrent O_DIRECT + writes */ + OCFS2_MOUNT_HB_NONE = 1 << 13, /* No heartbeat */ + OCFS2_MOUNT_HB_GLOBAL = 1 << 14, /* Global heartbeat */ }; #define OCFS2_OSB_SOFT_RO 0x0001 @@ -277,7 +288,8 @@ struct ocfs2_super struct super_block *sb; struct inode *root_inode; struct inode *sys_root_inode; - struct inode *system_inodes[NUM_SYSTEM_INODES]; + struct inode *global_system_inodes[NUM_GLOBAL_SYSTEM_INODES]; + struct inode **local_system_inodes; struct ocfs2_slot_info *slot_info; @@ -368,6 +380,8 @@ struct ocfs2_super struct ocfs2_alloc_stats alloc_stats; char dev_str[20]; /* "major,minor" of the device */ + u8 osb_stackflags; + char osb_cluster_stack[OCFS2_STACK_LABEL_LEN + 1]; struct ocfs2_cluster_connection *cconn; struct ocfs2_lock_res osb_super_lockres; @@ -601,10 +615,35 @@ static inline int ocfs2_is_soft_readonly(struct ocfs2_super *osb) return ret; } -static inline int ocfs2_userspace_stack(struct ocfs2_super *osb) +static inline int ocfs2_clusterinfo_valid(struct ocfs2_super *osb) { return (osb->s_feature_incompat & - OCFS2_FEATURE_INCOMPAT_USERSPACE_STACK); + (OCFS2_FEATURE_INCOMPAT_USERSPACE_STACK | + OCFS2_FEATURE_INCOMPAT_CLUSTERINFO)); +} + +static inline int ocfs2_userspace_stack(struct ocfs2_super *osb) +{ + if (ocfs2_clusterinfo_valid(osb) && + memcmp(osb->osb_cluster_stack, OCFS2_CLASSIC_CLUSTER_STACK, + OCFS2_STACK_LABEL_LEN)) + return 1; + return 0; +} + +static inline int ocfs2_o2cb_stack(struct ocfs2_super *osb) +{ + if (ocfs2_clusterinfo_valid(osb) && + !memcmp(osb->osb_cluster_stack, OCFS2_CLASSIC_CLUSTER_STACK, + OCFS2_STACK_LABEL_LEN)) + return 1; + return 0; +} + +static inline int ocfs2_cluster_o2cb_global_heartbeat(struct ocfs2_super *osb) +{ + return ocfs2_o2cb_stack(osb) && + (osb->osb_stackflags & OCFS2_CLUSTER_O2CB_GLOBAL_HEARTBEAT); } static inline int ocfs2_mount_local(struct ocfs2_super *osb) diff --git a/fs/ocfs2/ocfs2_fs.h b/fs/ocfs2/ocfs2_fs.h index 33f1c9a8258..c2e4f8222e2 100644 --- a/fs/ocfs2/ocfs2_fs.h +++ b/fs/ocfs2/ocfs2_fs.h @@ -101,7 +101,8 @@ | OCFS2_FEATURE_INCOMPAT_META_ECC \ | OCFS2_FEATURE_INCOMPAT_INDEXED_DIRS \ | OCFS2_FEATURE_INCOMPAT_REFCOUNT_TREE \ - | OCFS2_FEATURE_INCOMPAT_DISCONTIG_BG) + | OCFS2_FEATURE_INCOMPAT_DISCONTIG_BG \ + | OCFS2_FEATURE_INCOMPAT_CLUSTERINFO) #define OCFS2_FEATURE_RO_COMPAT_SUPP (OCFS2_FEATURE_RO_COMPAT_UNWRITTEN \ | OCFS2_FEATURE_RO_COMPAT_USRQUOTA \ | OCFS2_FEATURE_RO_COMPAT_GRPQUOTA) @@ -170,6 +171,13 @@ #define OCFS2_FEATURE_INCOMPAT_DISCONTIG_BG 0x2000 /* + * Incompat bit to indicate useable clusterinfo with stackflags for all + * cluster stacks (userspace adnd o2cb). If this bit is set, + * INCOMPAT_USERSPACE_STACK becomes superfluous and thus should not be set. + */ +#define OCFS2_FEATURE_INCOMPAT_CLUSTERINFO 0x4000 + +/* * backup superblock flag is used to indicate that this volume * has backup superblocks. */ @@ -235,18 +243,31 @@ #define OCFS2_HAS_REFCOUNT_FL (0x0010) /* Inode attributes, keep in sync with EXT2 */ -#define OCFS2_SECRM_FL (0x00000001) /* Secure deletion */ -#define OCFS2_UNRM_FL (0x00000002) /* Undelete */ -#define OCFS2_COMPR_FL (0x00000004) /* Compress file */ -#define OCFS2_SYNC_FL (0x00000008) /* Synchronous updates */ -#define OCFS2_IMMUTABLE_FL (0x00000010) /* Immutable file */ -#define OCFS2_APPEND_FL (0x00000020) /* writes to file may only append */ -#define OCFS2_NODUMP_FL (0x00000040) /* do not dump file */ -#define OCFS2_NOATIME_FL (0x00000080) /* do not update atime */ -#define OCFS2_DIRSYNC_FL (0x00010000) /* dirsync behaviour (directories only) */ - -#define OCFS2_FL_VISIBLE (0x000100FF) /* User visible flags */ -#define OCFS2_FL_MODIFIABLE (0x000100FF) /* User modifiable flags */ +#define OCFS2_SECRM_FL FS_SECRM_FL /* Secure deletion */ +#define OCFS2_UNRM_FL FS_UNRM_FL /* Undelete */ +#define OCFS2_COMPR_FL FS_COMPR_FL /* Compress file */ +#define OCFS2_SYNC_FL FS_SYNC_FL /* Synchronous updates */ +#define OCFS2_IMMUTABLE_FL FS_IMMUTABLE_FL /* Immutable file */ +#define OCFS2_APPEND_FL FS_APPEND_FL /* writes to file may only append */ +#define OCFS2_NODUMP_FL FS_NODUMP_FL /* do not dump file */ +#define OCFS2_NOATIME_FL FS_NOATIME_FL /* do not update atime */ +/* Reserved for compression usage... */ +#define OCFS2_DIRTY_FL FS_DIRTY_FL +#define OCFS2_COMPRBLK_FL FS_COMPRBLK_FL /* One or more compressed clusters */ +#define OCFS2_NOCOMP_FL FS_NOCOMP_FL /* Don't compress */ +#define OCFS2_ECOMPR_FL FS_ECOMPR_FL /* Compression error */ +/* End compression flags --- maybe not all used */ +#define OCFS2_BTREE_FL FS_BTREE_FL /* btree format dir */ +#define OCFS2_INDEX_FL FS_INDEX_FL /* hash-indexed directory */ +#define OCFS2_IMAGIC_FL FS_IMAGIC_FL /* AFS directory */ +#define OCFS2_JOURNAL_DATA_FL FS_JOURNAL_DATA_FL /* Reserved for ext3 */ +#define OCFS2_NOTAIL_FL FS_NOTAIL_FL /* file tail should not be merged */ +#define OCFS2_DIRSYNC_FL FS_DIRSYNC_FL /* dirsync behaviour (directories only) */ +#define OCFS2_TOPDIR_FL FS_TOPDIR_FL /* Top of directory hierarchies*/ +#define OCFS2_RESERVED_FL FS_RESERVED_FL /* reserved for ext2 lib */ + +#define OCFS2_FL_VISIBLE FS_FL_USER_VISIBLE /* User visible flags */ +#define OCFS2_FL_MODIFIABLE FS_FL_USER_MODIFIABLE /* User modifiable flags */ /* * Extent record flags (e_node.leaf.flags) @@ -279,10 +300,13 @@ #define OCFS2_VOL_UUID_LEN 16 #define OCFS2_MAX_VOL_LABEL_LEN 64 -/* The alternate, userspace stack fields */ +/* The cluster stack fields */ #define OCFS2_STACK_LABEL_LEN 4 #define OCFS2_CLUSTER_NAME_LEN 16 +/* Classic (historically speaking) cluster stack */ +#define OCFS2_CLASSIC_CLUSTER_STACK "o2cb" + /* Journal limits (in bytes) */ #define OCFS2_MIN_JOURNAL_SIZE (4 * 1024 * 1024) @@ -292,6 +316,11 @@ */ #define OCFS2_MIN_XATTR_INLINE_SIZE 256 +/* + * Cluster info flags (ocfs2_cluster_info.ci_stackflags) + */ +#define OCFS2_CLUSTER_O2CB_GLOBAL_HEARTBEAT (0x01) + struct ocfs2_system_inode_info { char *si_name; int si_iflags; @@ -309,6 +338,7 @@ enum { USER_QUOTA_SYSTEM_INODE, GROUP_QUOTA_SYSTEM_INODE, #define OCFS2_LAST_GLOBAL_SYSTEM_INODE GROUP_QUOTA_SYSTEM_INODE +#define OCFS2_FIRST_LOCAL_SYSTEM_INODE ORPHAN_DIR_SYSTEM_INODE ORPHAN_DIR_SYSTEM_INODE, EXTENT_ALLOC_SYSTEM_INODE, INODE_ALLOC_SYSTEM_INODE, @@ -317,8 +347,12 @@ enum { TRUNCATE_LOG_SYSTEM_INODE, LOCAL_USER_QUOTA_SYSTEM_INODE, LOCAL_GROUP_QUOTA_SYSTEM_INODE, +#define OCFS2_LAST_LOCAL_SYSTEM_INODE LOCAL_GROUP_QUOTA_SYSTEM_INODE NUM_SYSTEM_INODES }; +#define NUM_GLOBAL_SYSTEM_INODES OCFS2_LAST_GLOBAL_SYSTEM_INODE +#define NUM_LOCAL_SYSTEM_INODES \ + (NUM_SYSTEM_INODES - OCFS2_FIRST_LOCAL_SYSTEM_INODE) static struct ocfs2_system_inode_info ocfs2_system_inodes[NUM_SYSTEM_INODES] = { /* Global system inodes (single copy) */ @@ -347,6 +381,7 @@ static struct ocfs2_system_inode_info ocfs2_system_inodes[NUM_SYSTEM_INODES] = { /* Parameter passed from mount.ocfs2 to module */ #define OCFS2_HB_NONE "heartbeat=none" #define OCFS2_HB_LOCAL "heartbeat=local" +#define OCFS2_HB_GLOBAL "heartbeat=global" /* * OCFS2 directory file types. Only the low 3 bits are used. The @@ -553,9 +588,21 @@ struct ocfs2_slot_map_extended { */ }; +/* + * ci_stackflags is only valid if the incompat bit + * OCFS2_FEATURE_INCOMPAT_CLUSTERINFO is set. + */ struct ocfs2_cluster_info { /*00*/ __u8 ci_stack[OCFS2_STACK_LABEL_LEN]; - __le32 ci_reserved; + union { + __le32 ci_reserved; + struct { + __u8 ci_stackflags; + __u8 ci_reserved1; + __u8 ci_reserved2; + __u8 ci_reserved3; + }; + }; /*08*/ __u8 ci_cluster[OCFS2_CLUSTER_NAME_LEN]; /*18*/ }; @@ -592,9 +639,9 @@ struct ocfs2_super_block { * group header */ /*50*/ __u8 s_label[OCFS2_MAX_VOL_LABEL_LEN]; /* Label for mounting, etc. */ /*90*/ __u8 s_uuid[OCFS2_VOL_UUID_LEN]; /* 128-bit uuid */ -/*A0*/ struct ocfs2_cluster_info s_cluster_info; /* Selected userspace - stack. Only valid - with INCOMPAT flag. */ +/*A0*/ struct ocfs2_cluster_info s_cluster_info; /* Only valid if either + userspace or clusterinfo + INCOMPAT flag set. */ /*B8*/ __le16 s_xattr_inline_size; /* extended attribute inline size for this fs*/ __le16 s_reserved0; diff --git a/fs/ocfs2/ocfs2_ioctl.h b/fs/ocfs2/ocfs2_ioctl.h index 2d3420af1a8..b46f39bf743 100644 --- a/fs/ocfs2/ocfs2_ioctl.h +++ b/fs/ocfs2/ocfs2_ioctl.h @@ -23,10 +23,10 @@ /* * ioctl commands */ -#define OCFS2_IOC_GETFLAGS _IOR('f', 1, long) -#define OCFS2_IOC_SETFLAGS _IOW('f', 2, long) -#define OCFS2_IOC32_GETFLAGS _IOR('f', 1, int) -#define OCFS2_IOC32_SETFLAGS _IOW('f', 2, int) +#define OCFS2_IOC_GETFLAGS FS_IOC_GETFLAGS +#define OCFS2_IOC_SETFLAGS FS_IOC_SETFLAGS +#define OCFS2_IOC32_GETFLAGS FS_IOC32_GETFLAGS +#define OCFS2_IOC32_SETFLAGS FS_IOC32_SETFLAGS /* * Space reservation / allocation / free ioctls and argument structure @@ -76,4 +76,99 @@ struct reflink_arguments { }; #define OCFS2_IOC_REFLINK _IOW('o', 4, struct reflink_arguments) +/* Following definitions dedicated for ocfs2_info_request ioctls. */ +#define OCFS2_INFO_MAX_REQUEST (50) +#define OCFS2_TEXT_UUID_LEN (OCFS2_VOL_UUID_LEN * 2) + +/* Magic number of all requests */ +#define OCFS2_INFO_MAGIC (0x4F32494E) + +/* + * Always try to separate info request into small pieces to + * guarantee the backward&forward compatibility. + */ +struct ocfs2_info { + __u64 oi_requests; /* Array of __u64 pointers to requests */ + __u32 oi_count; /* Number of requests in info_requests */ + __u32 oi_pad; +}; + +struct ocfs2_info_request { +/*00*/ __u32 ir_magic; /* Magic number */ + __u32 ir_code; /* Info request code */ + __u32 ir_size; /* Size of request */ + __u32 ir_flags; /* Request flags */ +/*10*/ /* Request specific fields */ +}; + +struct ocfs2_info_clustersize { + struct ocfs2_info_request ic_req; + __u32 ic_clustersize; + __u32 ic_pad; +}; + +struct ocfs2_info_blocksize { + struct ocfs2_info_request ib_req; + __u32 ib_blocksize; + __u32 ib_pad; +}; + +struct ocfs2_info_maxslots { + struct ocfs2_info_request im_req; + __u32 im_max_slots; + __u32 im_pad; +}; + +struct ocfs2_info_label { + struct ocfs2_info_request il_req; + __u8 il_label[OCFS2_MAX_VOL_LABEL_LEN]; +} __attribute__ ((packed)); + +struct ocfs2_info_uuid { + struct ocfs2_info_request iu_req; + __u8 iu_uuid_str[OCFS2_TEXT_UUID_LEN + 1]; +} __attribute__ ((packed)); + +struct ocfs2_info_fs_features { + struct ocfs2_info_request if_req; + __u32 if_compat_features; + __u32 if_incompat_features; + __u32 if_ro_compat_features; + __u32 if_pad; +}; + +struct ocfs2_info_journal_size { + struct ocfs2_info_request ij_req; + __u64 ij_journal_size; +}; + +/* Codes for ocfs2_info_request */ +enum ocfs2_info_type { + OCFS2_INFO_CLUSTERSIZE = 1, + OCFS2_INFO_BLOCKSIZE, + OCFS2_INFO_MAXSLOTS, + OCFS2_INFO_LABEL, + OCFS2_INFO_UUID, + OCFS2_INFO_FS_FEATURES, + OCFS2_INFO_JOURNAL_SIZE, + OCFS2_INFO_NUM_TYPES +}; + +/* Flags for struct ocfs2_info_request */ +/* Filled by the caller */ +#define OCFS2_INFO_FL_NON_COHERENT (0x00000001) /* Cluster coherency not + required. This is a hint. + It is up to ocfs2 whether + the request can be fulfilled + without locking. */ +/* Filled by ocfs2 */ +#define OCFS2_INFO_FL_FILLED (0x40000000) /* Filesystem understood + this request and + filled in the answer */ + +#define OCFS2_INFO_FL_ERROR (0x80000000) /* Error happened during + request handling. */ + +#define OCFS2_IOC_INFO _IOR('o', 5, struct ocfs2_info) + #endif /* OCFS2_IOCTL_H */ diff --git a/fs/ocfs2/refcounttree.c b/fs/ocfs2/refcounttree.c index 3ac5aa733e9..b5f9160e93e 100644 --- a/fs/ocfs2/refcounttree.c +++ b/fs/ocfs2/refcounttree.c @@ -49,6 +49,7 @@ struct ocfs2_cow_context { struct inode *inode; + struct file *file; u32 cow_start; u32 cow_len; struct ocfs2_extent_tree data_et; @@ -2436,16 +2437,26 @@ static int ocfs2_calc_refcount_meta_credits(struct super_block *sb, len = min((u64)cpos + clusters, le64_to_cpu(rec.r_cpos) + le32_to_cpu(rec.r_clusters)) - cpos; /* - * If the refcount rec already exist, cool. We just need - * to check whether there is a split. Otherwise we just need - * to increase the refcount. - * If we will insert one, increases recs_add. - * * We record all the records which will be inserted to the * same refcount block, so that we can tell exactly whether * we need a new refcount block or not. + * + * If we will insert a new one, this is easy and only happens + * during adding refcounted flag to the extent, so we don't + * have a chance of spliting. We just need one record. + * + * If the refcount rec already exists, that would be a little + * complicated. we may have to: + * 1) split at the beginning if the start pos isn't aligned. + * we need 1 more record in this case. + * 2) split int the end if the end pos isn't aligned. + * we need 1 more record in this case. + * 3) split in the middle because of file system fragmentation. + * we need 2 more records in this case(we can't detect this + * beforehand, so always think of the worst case). */ if (rec.r_refcount) { + recs_add += 2; /* Check whether we need a split at the beginning. */ if (cpos == start_cpos && cpos != le64_to_cpu(rec.r_cpos)) @@ -2922,13 +2933,16 @@ static int ocfs2_duplicate_clusters_by_page(handle_t *handle, u64 new_block = ocfs2_clusters_to_blocks(sb, new_cluster); struct page *page; pgoff_t page_index; - unsigned int from, to; + unsigned int from, to, readahead_pages; loff_t offset, end, map_end; struct address_space *mapping = context->inode->i_mapping; mlog(0, "old_cluster %u, new %u, len %u at offset %u\n", old_cluster, new_cluster, new_len, cpos); + readahead_pages = + (ocfs2_cow_contig_clusters(sb) << + OCFS2_SB(sb)->s_clustersize_bits) >> PAGE_CACHE_SHIFT; offset = ((loff_t)cpos) << OCFS2_SB(sb)->s_clustersize_bits; end = offset + (new_len << OCFS2_SB(sb)->s_clustersize_bits); /* @@ -2950,7 +2964,7 @@ static int ocfs2_duplicate_clusters_by_page(handle_t *handle, if (map_end & (PAGE_CACHE_SIZE - 1)) to = map_end & (PAGE_CACHE_SIZE - 1); - page = grab_cache_page(mapping, page_index); + page = find_or_create_page(mapping, page_index, GFP_NOFS); /* * In case PAGE_CACHE_SIZE <= CLUSTER_SIZE, This page @@ -2959,6 +2973,14 @@ 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) { @@ -3169,7 +3191,8 @@ static int ocfs2_cow_sync_writeback(struct super_block *sb, if (map_end > end) map_end = end; - page = grab_cache_page(context->inode->i_mapping, page_index); + page = find_or_create_page(context->inode->i_mapping, + page_index, GFP_NOFS); BUG_ON(!page); wait_on_page_writeback(page); @@ -3398,12 +3421,35 @@ 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) { @@ -3432,6 +3478,8 @@ 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; @@ -3453,6 +3501,7 @@ 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); @@ -3481,6 +3530,7 @@ 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) { @@ -3500,7 +3550,7 @@ int ocfs2_refcount_cow(struct inode *inode, num_clusters = write_len; if (ext_flags & OCFS2_EXT_REFCOUNTED) { - ret = ocfs2_refcount_cow_hunk(inode, di_bh, cpos, + ret = ocfs2_refcount_cow_hunk(inode, file, di_bh, cpos, num_clusters, max_cpos); if (ret) { mlog_errno(ret); @@ -4190,8 +4240,9 @@ static int __ocfs2_reflink(struct dentry *old_dentry, goto out; } - mutex_lock(&new_inode->i_mutex); - ret = ocfs2_inode_lock(new_inode, &new_bh, 1); + mutex_lock_nested(&new_inode->i_mutex, I_MUTEX_CHILD); + ret = ocfs2_inode_lock_nested(new_inode, &new_bh, 1, + OI_LS_REFLINK_TARGET); if (ret) { mlog_errno(ret); goto out_unlock; diff --git a/fs/ocfs2/refcounttree.h b/fs/ocfs2/refcounttree.h index 9983ba1570e..c8ce46f7d8e 100644 --- a/fs/ocfs2/refcounttree.h +++ b/fs/ocfs2/refcounttree.h @@ -21,14 +21,14 @@ struct ocfs2_refcount_tree { struct rb_node rf_node; u64 rf_blkno; u32 rf_generation; + struct kref rf_getcnt; struct rw_semaphore rf_sem; struct ocfs2_lock_res rf_lockres; - struct kref rf_getcnt; int rf_removed; /* the following 4 fields are used by caching_info. */ - struct ocfs2_caching_info rf_ci; spinlock_t rf_lock; + struct ocfs2_caching_info rf_ci; struct mutex rf_io_mutex; struct super_block *rf_sb; }; @@ -52,7 +52,8 @@ int ocfs2_prepare_refcount_change_for_del(struct inode *inode, u32 clusters, int *credits, int *ref_blocks); -int ocfs2_refcount_cow(struct inode *inode, struct buffer_head *di_bh, +int ocfs2_refcount_cow(struct inode *inode, + struct file *filep, struct buffer_head *di_bh, u32 cpos, u32 write_len, u32 max_cpos); typedef int (ocfs2_post_refcount_func)(struct inode *inode, diff --git a/fs/ocfs2/reservations.c b/fs/ocfs2/reservations.c index d8b6e4259b8..3e78db361bc 100644 --- a/fs/ocfs2/reservations.c +++ b/fs/ocfs2/reservations.c @@ -732,25 +732,23 @@ int ocfs2_resmap_resv_bits(struct ocfs2_reservation_map *resmap, struct ocfs2_alloc_reservation *resv, int *cstart, int *clen) { - unsigned int wanted = *clen; - if (resv == NULL || ocfs2_resmap_disabled(resmap)) return -ENOSPC; spin_lock(&resv_lock); - /* - * We don't want to over-allocate for temporary - * windows. Otherwise, we run the risk of fragmenting the - * allocation space. - */ - wanted = ocfs2_resv_window_bits(resmap, resv); - if ((resv->r_flags & OCFS2_RESV_FLAG_TMP) || wanted < *clen) - wanted = *clen; - if (ocfs2_resv_empty(resv)) { - mlog(0, "empty reservation, find new window\n"); + /* + * We don't want to over-allocate for temporary + * windows. Otherwise, we run the risk of fragmenting the + * allocation space. + */ + unsigned int wanted = ocfs2_resv_window_bits(resmap, resv); + if ((resv->r_flags & OCFS2_RESV_FLAG_TMP) || wanted < *clen) + wanted = *clen; + + mlog(0, "empty reservation, find new window\n"); /* * Try to get a window here. If it works, we must fall * through and test the bitmap . This avoids some diff --git a/fs/ocfs2/slot_map.c b/fs/ocfs2/slot_map.c index bfbd7e9e949..ab4e0172cc1 100644 --- a/fs/ocfs2/slot_map.c +++ b/fs/ocfs2/slot_map.c @@ -357,7 +357,7 @@ static int ocfs2_map_slot_buffers(struct ocfs2_super *osb, { int status = 0; u64 blkno; - unsigned long long blocks, bytes; + unsigned long long blocks, bytes = 0; unsigned int i; struct buffer_head *bh; diff --git a/fs/ocfs2/stack_o2cb.c b/fs/ocfs2/stack_o2cb.c index 0d3049f696c..19965b00c43 100644 --- a/fs/ocfs2/stack_o2cb.c +++ b/fs/ocfs2/stack_o2cb.c @@ -283,6 +283,8 @@ static int o2cb_cluster_connect(struct ocfs2_cluster_connection *conn) /* 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; goto out; } diff --git a/fs/ocfs2/stack_user.c b/fs/ocfs2/stack_user.c index 2dc57bca068..252e7c82f92 100644 --- a/fs/ocfs2/stack_user.c +++ b/fs/ocfs2/stack_user.c @@ -22,7 +22,6 @@ #include <linux/miscdevice.h> #include <linux/mutex.h> #include <linux/slab.h> -#include <linux/smp_lock.h> #include <linux/reboot.h> #include <asm/uaccess.h> @@ -612,12 +611,10 @@ static int ocfs2_control_open(struct inode *inode, struct file *file) return -ENOMEM; p->op_this_node = -1; - lock_kernel(); mutex_lock(&ocfs2_control_lock); file->private_data = p; list_add(&p->op_list, &ocfs2_control_private_list); mutex_unlock(&ocfs2_control_lock); - unlock_kernel(); return 0; } @@ -628,6 +625,7 @@ static const struct file_operations ocfs2_control_fops = { .read = ocfs2_control_read, .write = ocfs2_control_write, .owner = THIS_MODULE, + .llseek = default_llseek, }; static struct miscdevice ocfs2_control_device = { diff --git a/fs/ocfs2/suballoc.c b/fs/ocfs2/suballoc.c index a8e6a95a353..5fed60de763 100644 --- a/fs/ocfs2/suballoc.c +++ b/fs/ocfs2/suballoc.c @@ -57,11 +57,28 @@ struct ocfs2_suballoc_result { u64 sr_bg_blkno; /* The bg we allocated from. Set to 0 when a block group is contiguous. */ + u64 sr_bg_stable_blkno; /* + * Doesn't change, always + * set to target block + * group descriptor + * block. + */ u64 sr_blkno; /* The first allocated block */ unsigned int sr_bit_offset; /* The bit in the bg */ unsigned int sr_bits; /* How many bits we claimed */ }; +static u64 ocfs2_group_from_res(struct ocfs2_suballoc_result *res) +{ + if (res->sr_blkno == 0) + return 0; + + if (res->sr_bg_blkno) + return res->sr_bg_blkno; + + return ocfs2_which_suballoc_group(res->sr_blkno, res->sr_bit_offset); +} + static inline void ocfs2_debug_bg(struct ocfs2_group_desc *bg); static inline void ocfs2_debug_suballoc_inode(struct ocfs2_dinode *fe); static inline u16 ocfs2_find_victim_chain(struct ocfs2_chain_list *cl); @@ -138,6 +155,10 @@ void ocfs2_free_ac_resource(struct ocfs2_alloc_context *ac) brelse(ac->ac_bh); ac->ac_bh = NULL; ac->ac_resv = NULL; + if (ac->ac_find_loc_priv) { + kfree(ac->ac_find_loc_priv); + ac->ac_find_loc_priv = NULL; + } } void ocfs2_free_alloc_context(struct ocfs2_alloc_context *ac) @@ -336,7 +357,7 @@ out: static void ocfs2_bg_discontig_add_extent(struct ocfs2_super *osb, struct ocfs2_group_desc *bg, struct ocfs2_chain_list *cl, - u64 p_blkno, u32 clusters) + u64 p_blkno, unsigned int clusters) { struct ocfs2_extent_list *el = &bg->bg_list; struct ocfs2_extent_rec *rec; @@ -348,7 +369,7 @@ static void ocfs2_bg_discontig_add_extent(struct ocfs2_super *osb, rec->e_blkno = cpu_to_le64(p_blkno); rec->e_cpos = cpu_to_le32(le16_to_cpu(bg->bg_bits) / le16_to_cpu(cl->cl_bpc)); - rec->e_leaf_clusters = cpu_to_le32(clusters); + rec->e_leaf_clusters = cpu_to_le16(clusters); le16_add_cpu(&bg->bg_bits, clusters * le16_to_cpu(cl->cl_bpc)); le16_add_cpu(&bg->bg_free_bits_count, clusters * le16_to_cpu(cl->cl_bpc)); @@ -1359,6 +1380,14 @@ static inline int ocfs2_block_group_set_bits(handle_t *handle, } le16_add_cpu(&bg->bg_free_bits_count, -num_bits); + if (le16_to_cpu(bg->bg_free_bits_count) > le16_to_cpu(bg->bg_bits)) { + ocfs2_error(alloc_inode->i_sb, "Group descriptor # %llu has bit" + " count %u but claims %u are freed. num_bits %d", + (unsigned long long)le64_to_cpu(bg->bg_blkno), + le16_to_cpu(bg->bg_bits), + le16_to_cpu(bg->bg_free_bits_count), num_bits); + return -EROFS; + } while(num_bits--) ocfs2_set_bit(bit_off++, bitmap); @@ -1678,6 +1707,15 @@ static int ocfs2_search_one_group(struct ocfs2_alloc_context *ac, if (!ret) ocfs2_bg_discontig_fix_result(ac, gd, res); + /* + * sr_bg_blkno might have been changed by + * ocfs2_bg_discontig_fix_result + */ + res->sr_bg_stable_blkno = group_bh->b_blocknr; + + if (ac->ac_find_loc_only) + goto out_loc_only; + ret = ocfs2_alloc_dinode_update_counts(alloc_inode, handle, ac->ac_bh, res->sr_bits, le16_to_cpu(gd->bg_chain)); @@ -1691,6 +1729,7 @@ static int ocfs2_search_one_group(struct ocfs2_alloc_context *ac, if (ret < 0) mlog_errno(ret); +out_loc_only: *bits_left = le16_to_cpu(gd->bg_free_bits_count); out: @@ -1708,7 +1747,6 @@ static int ocfs2_search_chain(struct ocfs2_alloc_context *ac, { int status; u16 chain; - u32 tmp_used; u64 next_group; struct inode *alloc_inode = ac->ac_inode; struct buffer_head *group_bh = NULL; @@ -1770,6 +1808,11 @@ static int ocfs2_search_chain(struct ocfs2_alloc_context *ac, if (!status) ocfs2_bg_discontig_fix_result(ac, bg, res); + /* + * sr_bg_blkno might have been changed by + * ocfs2_bg_discontig_fix_result + */ + res->sr_bg_stable_blkno = group_bh->b_blocknr; /* * Keep track of previous block descriptor read. When @@ -1796,22 +1839,17 @@ static int ocfs2_search_chain(struct ocfs2_alloc_context *ac, } } - /* Ok, claim our bits now: set the info on dinode, chainlist - * and then the group */ - status = ocfs2_journal_access_di(handle, - INODE_CACHE(alloc_inode), - ac->ac_bh, - OCFS2_JOURNAL_ACCESS_WRITE); - if (status < 0) { + if (ac->ac_find_loc_only) + goto out_loc_only; + + status = ocfs2_alloc_dinode_update_counts(alloc_inode, handle, + ac->ac_bh, res->sr_bits, + chain); + if (status) { mlog_errno(status); goto bail; } - tmp_used = le32_to_cpu(fe->id1.bitmap1.i_used); - fe->id1.bitmap1.i_used = cpu_to_le32(res->sr_bits + tmp_used); - le32_add_cpu(&cl->cl_recs[chain].c_free, -res->sr_bits); - ocfs2_journal_dirty(handle, ac->ac_bh); - status = ocfs2_block_group_set_bits(handle, alloc_inode, bg, @@ -1826,6 +1864,7 @@ static int ocfs2_search_chain(struct ocfs2_alloc_context *ac, mlog(0, "Allocated %u bits from suballocator %llu\n", res->sr_bits, (unsigned long long)le64_to_cpu(fe->i_blkno)); +out_loc_only: *bits_left = le16_to_cpu(bg->bg_free_bits_count); bail: brelse(group_bh); @@ -1845,6 +1884,7 @@ static int ocfs2_claim_suballoc_bits(struct ocfs2_alloc_context *ac, int status; u16 victim, i; u16 bits_left = 0; + u64 hint = ac->ac_last_group; struct ocfs2_chain_list *cl; struct ocfs2_dinode *fe; @@ -1872,7 +1912,7 @@ static int ocfs2_claim_suballoc_bits(struct ocfs2_alloc_context *ac, goto bail; } - res->sr_bg_blkno = ac->ac_last_group; + res->sr_bg_blkno = hint; if (res->sr_bg_blkno) { /* Attempt to short-circuit the usual search mechanism * by jumping straight to the most recently used @@ -1896,8 +1936,10 @@ static int ocfs2_claim_suballoc_bits(struct ocfs2_alloc_context *ac, status = ocfs2_search_chain(ac, handle, bits_wanted, min_bits, res, &bits_left); - if (!status) + if (!status) { + hint = ocfs2_group_from_res(res); goto set_hint; + } if (status < 0 && status != -ENOSPC) { mlog_errno(status); goto bail; @@ -1920,8 +1962,10 @@ static int ocfs2_claim_suballoc_bits(struct ocfs2_alloc_context *ac, ac->ac_chain = i; status = ocfs2_search_chain(ac, handle, bits_wanted, min_bits, res, &bits_left); - if (!status) + if (!status) { + hint = ocfs2_group_from_res(res); break; + } if (status < 0 && status != -ENOSPC) { mlog_errno(status); goto bail; @@ -1936,7 +1980,7 @@ set_hint: if (bits_left < min_bits) ac->ac_last_group = 0; else - ac->ac_last_group = res->sr_bg_blkno; + ac->ac_last_group = hint; } bail: @@ -2016,6 +2060,136 @@ static inline void ocfs2_save_inode_ac_group(struct inode *dir, OCFS2_I(dir)->ip_last_used_slot = ac->ac_alloc_slot; } +int ocfs2_find_new_inode_loc(struct inode *dir, + struct buffer_head *parent_fe_bh, + struct ocfs2_alloc_context *ac, + u64 *fe_blkno) +{ + int ret; + handle_t *handle = NULL; + struct ocfs2_suballoc_result *res; + + BUG_ON(!ac); + BUG_ON(ac->ac_bits_given != 0); + BUG_ON(ac->ac_bits_wanted != 1); + BUG_ON(ac->ac_which != OCFS2_AC_USE_INODE); + + res = kzalloc(sizeof(*res), GFP_NOFS); + if (res == NULL) { + ret = -ENOMEM; + mlog_errno(ret); + goto out; + } + + ocfs2_init_inode_ac_group(dir, parent_fe_bh, ac); + + /* + * The handle started here is for chain relink. Alternatively, + * we could just disable relink for these calls. + */ + handle = ocfs2_start_trans(OCFS2_SB(dir->i_sb), OCFS2_SUBALLOC_ALLOC); + if (IS_ERR(handle)) { + ret = PTR_ERR(handle); + handle = NULL; + mlog_errno(ret); + goto out; + } + + /* + * This will instruct ocfs2_claim_suballoc_bits and + * ocfs2_search_one_group to search but save actual allocation + * for later. + */ + ac->ac_find_loc_only = 1; + + ret = ocfs2_claim_suballoc_bits(ac, handle, 1, 1, res); + if (ret < 0) { + mlog_errno(ret); + goto out; + } + + ac->ac_find_loc_priv = res; + *fe_blkno = res->sr_blkno; + +out: + if (handle) + ocfs2_commit_trans(OCFS2_SB(dir->i_sb), handle); + + if (ret) + kfree(res); + + return ret; +} + +int ocfs2_claim_new_inode_at_loc(handle_t *handle, + struct inode *dir, + struct ocfs2_alloc_context *ac, + u64 *suballoc_loc, + u16 *suballoc_bit, + u64 di_blkno) +{ + int ret; + u16 chain; + struct ocfs2_suballoc_result *res = ac->ac_find_loc_priv; + struct buffer_head *bg_bh = NULL; + struct ocfs2_group_desc *bg; + struct ocfs2_dinode *di = (struct ocfs2_dinode *) ac->ac_bh->b_data; + + /* + * Since di_blkno is being passed back in, we check for any + * inconsistencies which may have happened between + * calls. These are code bugs as di_blkno is not expected to + * change once returned from ocfs2_find_new_inode_loc() + */ + BUG_ON(res->sr_blkno != di_blkno); + + ret = ocfs2_read_group_descriptor(ac->ac_inode, di, + res->sr_bg_stable_blkno, &bg_bh); + if (ret) { + mlog_errno(ret); + goto out; + } + + bg = (struct ocfs2_group_desc *) bg_bh->b_data; + chain = le16_to_cpu(bg->bg_chain); + + ret = ocfs2_alloc_dinode_update_counts(ac->ac_inode, handle, + ac->ac_bh, res->sr_bits, + chain); + if (ret) { + mlog_errno(ret); + goto out; + } + + ret = ocfs2_block_group_set_bits(handle, + ac->ac_inode, + bg, + bg_bh, + res->sr_bit_offset, + res->sr_bits); + if (ret < 0) { + mlog_errno(ret); + goto out; + } + + mlog(0, "Allocated %u bits from suballocator %llu\n", res->sr_bits, + (unsigned long long)di_blkno); + + atomic_inc(&OCFS2_SB(ac->ac_inode->i_sb)->alloc_stats.bg_allocs); + + BUG_ON(res->sr_bits != 1); + + *suballoc_loc = res->sr_bg_blkno; + *suballoc_bit = res->sr_bit_offset; + ac->ac_bits_given++; + ocfs2_save_inode_ac_group(dir, ac); + +out: + brelse(bg_bh); + + return ret; +} + int ocfs2_claim_new_inode(handle_t *handle, struct inode *dir, struct buffer_head *parent_fe_bh, @@ -2253,6 +2427,14 @@ static int ocfs2_block_group_clear_bits(handle_t *handle, (unsigned long *) undo_bg->bg_bitmap); } le16_add_cpu(&bg->bg_free_bits_count, num_bits); + if (le16_to_cpu(bg->bg_free_bits_count) > le16_to_cpu(bg->bg_bits)) { + ocfs2_error(alloc_inode->i_sb, "Group descriptor # %llu has bit" + " count %u but claims %u are freed. num_bits %d", + (unsigned long long)le64_to_cpu(bg->bg_blkno), + le16_to_cpu(bg->bg_bits), + le16_to_cpu(bg->bg_free_bits_count), num_bits); + return -EROFS; + } if (undo_fn) jbd_unlock_bh_state(group_bh); @@ -2567,7 +2749,8 @@ out: * suballoc_bit. */ static int ocfs2_get_suballoc_slot_bit(struct ocfs2_super *osb, u64 blkno, - u16 *suballoc_slot, u16 *suballoc_bit) + u16 *suballoc_slot, u64 *group_blkno, + u16 *suballoc_bit) { int status; struct buffer_head *inode_bh = NULL; @@ -2604,6 +2787,8 @@ static int ocfs2_get_suballoc_slot_bit(struct ocfs2_super *osb, u64 blkno, *suballoc_slot = le16_to_cpu(inode_fe->i_suballoc_slot); if (suballoc_bit) *suballoc_bit = le16_to_cpu(inode_fe->i_suballoc_bit); + if (group_blkno) + *group_blkno = le64_to_cpu(inode_fe->i_suballoc_loc); bail: brelse(inode_bh); @@ -2621,7 +2806,8 @@ bail: */ static int ocfs2_test_suballoc_bit(struct ocfs2_super *osb, struct inode *suballoc, - struct buffer_head *alloc_bh, u64 blkno, + struct buffer_head *alloc_bh, + u64 group_blkno, u64 blkno, u16 bit, int *res) { struct ocfs2_dinode *alloc_di; @@ -2642,10 +2828,8 @@ static int ocfs2_test_suballoc_bit(struct ocfs2_super *osb, goto bail; } - if (alloc_di->i_suballoc_loc) - bg_blkno = le64_to_cpu(alloc_di->i_suballoc_loc); - else - bg_blkno = ocfs2_which_suballoc_group(blkno, bit); + bg_blkno = group_blkno ? group_blkno : + ocfs2_which_suballoc_group(blkno, bit); status = ocfs2_read_group_descriptor(suballoc, alloc_di, bg_blkno, &group_bh); if (status < 0) { @@ -2680,6 +2864,7 @@ bail: int ocfs2_test_inode_bit(struct ocfs2_super *osb, u64 blkno, int *res) { int status; + u64 group_blkno = 0; u16 suballoc_bit = 0, suballoc_slot = 0; struct inode *inode_alloc_inode; struct buffer_head *alloc_bh = NULL; @@ -2687,7 +2872,7 @@ int ocfs2_test_inode_bit(struct ocfs2_super *osb, u64 blkno, int *res) mlog_entry("blkno: %llu", (unsigned long long)blkno); status = ocfs2_get_suballoc_slot_bit(osb, blkno, &suballoc_slot, - &suballoc_bit); + &group_blkno, &suballoc_bit); if (status < 0) { mlog(ML_ERROR, "get alloc slot and bit failed %d\n", status); goto bail; @@ -2715,7 +2900,7 @@ int ocfs2_test_inode_bit(struct ocfs2_super *osb, u64 blkno, int *res) } status = ocfs2_test_suballoc_bit(osb, inode_alloc_inode, alloc_bh, - blkno, suballoc_bit, res); + group_blkno, blkno, suballoc_bit, res); if (status < 0) mlog(ML_ERROR, "test suballoc bit failed %d\n", status); diff --git a/fs/ocfs2/suballoc.h b/fs/ocfs2/suballoc.h index a017dd3ee7d..b8afabfeede 100644 --- a/fs/ocfs2/suballoc.h +++ b/fs/ocfs2/suballoc.h @@ -56,6 +56,9 @@ struct ocfs2_alloc_context { u64 ac_max_block; /* Highest block number to allocate. 0 is is the same as ~0 - unlimited */ + int ac_find_loc_only; /* hack for reflink operation ordering */ + struct ocfs2_suballoc_result *ac_find_loc_priv; /* */ + struct ocfs2_alloc_reservation *ac_resv; }; @@ -197,4 +200,22 @@ int ocfs2_lock_allocators(struct inode *inode, struct ocfs2_extent_tree *et, struct ocfs2_alloc_context **meta_ac); int ocfs2_test_inode_bit(struct ocfs2_super *osb, u64 blkno, int *res); + + + +/* + * The following two interfaces are for ocfs2_create_inode_in_orphan(). + */ +int ocfs2_find_new_inode_loc(struct inode *dir, + struct buffer_head *parent_fe_bh, + struct ocfs2_alloc_context *ac, + u64 *fe_blkno); + +int ocfs2_claim_new_inode_at_loc(handle_t *handle, + struct inode *dir, + struct ocfs2_alloc_context *ac, + u64 *suballoc_loc, + u16 *suballoc_bit, + u64 di_blkno); + #endif /* _CHAINALLOC_H_ */ diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c index fa1be1b304d..56f0cb39582 100644 --- a/fs/ocfs2/super.c +++ b/fs/ocfs2/super.c @@ -162,6 +162,7 @@ enum { Opt_nointr, Opt_hb_none, Opt_hb_local, + Opt_hb_global, Opt_data_ordered, Opt_data_writeback, Opt_atime_quantum, @@ -177,6 +178,8 @@ enum { Opt_noacl, Opt_usrquota, Opt_grpquota, + Opt_coherency_buffered, + Opt_coherency_full, Opt_resv_level, Opt_dir_resv_level, Opt_err, @@ -190,6 +193,7 @@ static const match_table_t tokens = { {Opt_nointr, "nointr"}, {Opt_hb_none, OCFS2_HB_NONE}, {Opt_hb_local, OCFS2_HB_LOCAL}, + {Opt_hb_global, OCFS2_HB_GLOBAL}, {Opt_data_ordered, "data=ordered"}, {Opt_data_writeback, "data=writeback"}, {Opt_atime_quantum, "atime_quantum=%u"}, @@ -205,6 +209,8 @@ static const match_table_t tokens = { {Opt_noacl, "noacl"}, {Opt_usrquota, "usrquota"}, {Opt_grpquota, "grpquota"}, + {Opt_coherency_buffered, "coherency=buffered"}, + {Opt_coherency_full, "coherency=full"}, {Opt_resv_level, "resv_level=%u"}, {Opt_dir_resv_level, "dir_resv_level=%u"}, {Opt_err, NULL} @@ -514,11 +520,11 @@ static void ocfs2_release_system_inodes(struct ocfs2_super *osb) mlog_entry_void(); - for (i = 0; i < NUM_SYSTEM_INODES; i++) { - inode = osb->system_inodes[i]; + for (i = 0; i < NUM_GLOBAL_SYSTEM_INODES; i++) { + inode = osb->global_system_inodes[i]; if (inode) { iput(inode); - osb->system_inodes[i] = NULL; + osb->global_system_inodes[i] = NULL; } } @@ -534,6 +540,20 @@ static void ocfs2_release_system_inodes(struct ocfs2_super *osb) osb->root_inode = NULL; } + if (!osb->local_system_inodes) + goto out; + + for (i = 0; i < NUM_LOCAL_SYSTEM_INODES * osb->max_slots; i++) { + if (osb->local_system_inodes[i]) { + iput(osb->local_system_inodes[i]); + osb->local_system_inodes[i] = NULL; + } + } + + kfree(osb->local_system_inodes); + osb->local_system_inodes = NULL; + +out: mlog_exit(0); } @@ -608,8 +628,7 @@ static int ocfs2_remount(struct super_block *sb, int *flags, char *data) int ret = 0; struct mount_options parsed_options; struct ocfs2_super *osb = OCFS2_SB(sb); - - lock_kernel(); + u32 tmp; if (!ocfs2_parse_options(sb, data, &parsed_options, 1) || !ocfs2_check_set_options(sb, &parsed_options)) { @@ -617,8 +636,9 @@ static int ocfs2_remount(struct super_block *sb, int *flags, char *data) goto out; } - if ((osb->s_mount_opt & OCFS2_MOUNT_HB_LOCAL) != - (parsed_options.mount_opt & OCFS2_MOUNT_HB_LOCAL)) { + tmp = OCFS2_MOUNT_HB_LOCAL | OCFS2_MOUNT_HB_GLOBAL | + OCFS2_MOUNT_HB_NONE; + if ((osb->s_mount_opt & tmp) != (parsed_options.mount_opt & tmp)) { ret = -EINVAL; mlog(ML_ERROR, "Cannot change heartbeat mode on remount\n"); goto out; @@ -717,7 +737,6 @@ unlock_osb: MS_POSIXACL : 0); } out: - unlock_kernel(); return ret; } @@ -809,23 +828,29 @@ bail: static int ocfs2_verify_heartbeat(struct ocfs2_super *osb) { - if (ocfs2_mount_local(osb)) { - if (osb->s_mount_opt & OCFS2_MOUNT_HB_LOCAL) { + u32 hb_enabled = OCFS2_MOUNT_HB_LOCAL | OCFS2_MOUNT_HB_GLOBAL; + + if (osb->s_mount_opt & hb_enabled) { + if (ocfs2_mount_local(osb)) { mlog(ML_ERROR, "Cannot heartbeat on a locally " "mounted device.\n"); return -EINVAL; } - } - - if (ocfs2_userspace_stack(osb)) { - if (osb->s_mount_opt & OCFS2_MOUNT_HB_LOCAL) { + if (ocfs2_userspace_stack(osb)) { mlog(ML_ERROR, "Userspace stack expected, but " "o2cb heartbeat arguments passed to mount\n"); return -EINVAL; } + if (((osb->s_mount_opt & OCFS2_MOUNT_HB_GLOBAL) && + !ocfs2_cluster_o2cb_global_heartbeat(osb)) || + ((osb->s_mount_opt & OCFS2_MOUNT_HB_LOCAL) && + ocfs2_cluster_o2cb_global_heartbeat(osb))) { + mlog(ML_ERROR, "Mismatching o2cb heartbeat modes\n"); + return -EINVAL; + } } - if (!(osb->s_mount_opt & OCFS2_MOUNT_HB_LOCAL)) { + if (!(osb->s_mount_opt & hb_enabled)) { if (!ocfs2_mount_local(osb) && !ocfs2_is_hard_readonly(osb) && !ocfs2_userspace_stack(osb)) { mlog(ML_ERROR, "Heartbeat has to be started to mount " @@ -1291,6 +1316,7 @@ static int ocfs2_parse_options(struct super_block *sb, { int status; char *p; + u32 tmp; mlog_entry("remount: %d, options: \"%s\"\n", is_remount, options ? options : "(none)"); @@ -1322,7 +1348,10 @@ static int ocfs2_parse_options(struct super_block *sb, mopt->mount_opt |= OCFS2_MOUNT_HB_LOCAL; break; case Opt_hb_none: - mopt->mount_opt &= ~OCFS2_MOUNT_HB_LOCAL; + mopt->mount_opt |= OCFS2_MOUNT_HB_NONE; + break; + case Opt_hb_global: + mopt->mount_opt |= OCFS2_MOUNT_HB_GLOBAL; break; case Opt_barrier: if (match_int(&args[0], &option)) { @@ -1438,6 +1467,12 @@ static int ocfs2_parse_options(struct super_block *sb, case Opt_grpquota: mopt->mount_opt |= OCFS2_MOUNT_GRPQUOTA; break; + case Opt_coherency_buffered: + mopt->mount_opt |= OCFS2_MOUNT_COHERENCY_BUFFERED; + break; + case Opt_coherency_full: + mopt->mount_opt &= ~OCFS2_MOUNT_COHERENCY_BUFFERED; + break; case Opt_acl: mopt->mount_opt |= OCFS2_MOUNT_POSIX_ACL; mopt->mount_opt &= ~OCFS2_MOUNT_NO_POSIX_ACL; @@ -1477,6 +1512,15 @@ static int ocfs2_parse_options(struct super_block *sb, } } + /* Ensure only one heartbeat mode */ + tmp = mopt->mount_opt & (OCFS2_MOUNT_HB_LOCAL | OCFS2_MOUNT_HB_GLOBAL | + OCFS2_MOUNT_HB_NONE); + if (hweight32(tmp) != 1) { + mlog(ML_ERROR, "Invalid heartbeat mount options\n"); + status = 0; + goto bail; + } + status = 1; bail: @@ -1490,10 +1534,14 @@ static int ocfs2_show_options(struct seq_file *s, struct vfsmount *mnt) unsigned long opts = osb->s_mount_opt; unsigned int local_alloc_megs; - if (opts & OCFS2_MOUNT_HB_LOCAL) - seq_printf(s, ",_netdev,heartbeat=local"); - else - seq_printf(s, ",heartbeat=none"); + if (opts & (OCFS2_MOUNT_HB_LOCAL | OCFS2_MOUNT_HB_GLOBAL)) { + seq_printf(s, ",_netdev"); + if (opts & OCFS2_MOUNT_HB_LOCAL) + seq_printf(s, ",%s", OCFS2_HB_LOCAL); + else + seq_printf(s, ",%s", OCFS2_HB_GLOBAL); + } else + seq_printf(s, ",%s", OCFS2_HB_NONE); if (opts & OCFS2_MOUNT_NOINTR) seq_printf(s, ",nointr"); @@ -1536,6 +1584,11 @@ static int ocfs2_show_options(struct seq_file *s, struct vfsmount *mnt) if (opts & OCFS2_MOUNT_GRPQUOTA) seq_printf(s, ",grpquota"); + if (opts & OCFS2_MOUNT_COHERENCY_BUFFERED) + seq_printf(s, ",coherency=buffered"); + else + seq_printf(s, ",coherency=full"); + if (opts & OCFS2_MOUNT_NOUSERXATTR) seq_printf(s, ",nouser_xattr"); else @@ -1640,13 +1693,9 @@ static void ocfs2_put_super(struct super_block *sb) { mlog_entry("(0x%p)\n", sb); - lock_kernel(); - ocfs2_sync_blockdev(sb); ocfs2_dismount_volume(sb, 0); - unlock_kernel(); - mlog_exit_void(); } @@ -1990,6 +2039,36 @@ static int ocfs2_setup_osb_uuid(struct ocfs2_super *osb, const unsigned char *uu return 0; } +/* Make sure entire volume is addressable by our journal. Requires + osb_clusters_at_boot to be valid and for the journal to have been + initialized by ocfs2_journal_init(). */ +static int ocfs2_journal_addressable(struct ocfs2_super *osb) +{ + int status = 0; + u64 max_block = + ocfs2_clusters_to_blocks(osb->sb, + osb->osb_clusters_at_boot) - 1; + + /* 32-bit block number is always OK. */ + if (max_block <= (u32)~0ULL) + goto out; + + /* Volume is "huge", so see if our journal is new enough to + support it. */ + if (!(OCFS2_HAS_COMPAT_FEATURE(osb->sb, + OCFS2_FEATURE_COMPAT_JBD2_SB) && + jbd2_journal_check_used_features(osb->journal->j_journal, 0, 0, + JBD2_FEATURE_INCOMPAT_64BIT))) { + mlog(ML_ERROR, "The journal cannot address the entire volume. " + "Enable the 'block64' journal option with tunefs.ocfs2"); + status = -EFBIG; + goto out; + } + + out: + return status; +} + static int ocfs2_initialize_super(struct super_block *sb, struct buffer_head *bh, int sector_size, @@ -2002,6 +2081,7 @@ static int ocfs2_initialize_super(struct super_block *sb, struct ocfs2_journal *journal; __le32 uuid_net_key; struct ocfs2_super *osb; + u64 total_blocks; mlog_entry_void(); @@ -2060,6 +2140,15 @@ static int ocfs2_initialize_super(struct super_block *sb, snprintf(osb->dev_str, sizeof(osb->dev_str), "%u,%u", MAJOR(osb->sb->s_dev), MINOR(osb->sb->s_dev)); + osb->max_slots = le16_to_cpu(di->id2.i_super.s_max_slots); + if (osb->max_slots > OCFS2_MAX_SLOTS || osb->max_slots == 0) { + mlog(ML_ERROR, "Invalid number of node slots (%u)\n", + osb->max_slots); + status = -EINVAL; + goto bail; + } + mlog(0, "max_slots for this device: %u\n", osb->max_slots); + ocfs2_orphan_scan_init(osb); status = ocfs2_recovery_init(osb); @@ -2098,15 +2187,6 @@ static int ocfs2_initialize_super(struct super_block *sb, goto bail; } - osb->max_slots = le16_to_cpu(di->id2.i_super.s_max_slots); - if (osb->max_slots > OCFS2_MAX_SLOTS || osb->max_slots == 0) { - mlog(ML_ERROR, "Invalid number of node slots (%u)\n", - osb->max_slots); - status = -EINVAL; - goto bail; - } - mlog(0, "max_slots for this device: %u\n", osb->max_slots); - osb->slot_recovery_generations = kcalloc(osb->max_slots, sizeof(*osb->slot_recovery_generations), GFP_KERNEL); @@ -2149,7 +2229,9 @@ static int ocfs2_initialize_super(struct super_block *sb, goto bail; } - if (ocfs2_userspace_stack(osb)) { + if (ocfs2_clusterinfo_valid(osb)) { + osb->osb_stackflags = + OCFS2_RAW_SB(di)->s_cluster_info.ci_stackflags; memcpy(osb->osb_cluster_stack, OCFS2_RAW_SB(di)->s_cluster_info.ci_stack, OCFS2_STACK_LABEL_LEN); @@ -2214,11 +2296,15 @@ static int ocfs2_initialize_super(struct super_block *sb, goto bail; } - if (ocfs2_clusters_to_blocks(osb->sb, le32_to_cpu(di->i_clusters) - 1) - > (u32)~0UL) { - mlog(ML_ERROR, "Volume might try to write to blocks beyond " - "what jbd can address in 32 bits.\n"); - status = -EINVAL; + total_blocks = ocfs2_clusters_to_blocks(osb->sb, + le32_to_cpu(di->i_clusters)); + + status = generic_check_addressable(osb->sb->s_blocksize_bits, + total_blocks); + if (status) { + mlog(ML_ERROR, "Volume too large " + "to mount safely on this system"); + status = -EFBIG; goto bail; } @@ -2380,6 +2466,12 @@ static int ocfs2_check_volume(struct ocfs2_super *osb) goto finally; } + /* Now that journal has been initialized, check to make sure + entire volume is addressable. */ + status = ocfs2_journal_addressable(osb); + if (status) + goto finally; + /* If the journal was unmounted cleanly then we don't want to * recover anything. Otherwise, journal_load will do that * dirty work for us :) */ diff --git a/fs/ocfs2/symlink.c b/fs/ocfs2/symlink.c index 32499d213fc..9975457c981 100644 --- a/fs/ocfs2/symlink.c +++ b/fs/ocfs2/symlink.c @@ -128,7 +128,7 @@ static void *ocfs2_fast_follow_link(struct dentry *dentry, } /* Fast symlinks can't be large */ - len = strlen(target); + len = strnlen(target, ocfs2_fast_symlink_chars(inode->i_sb)); link = kzalloc(len + 1, GFP_NOFS); if (!link) { status = -ENOMEM; diff --git a/fs/ocfs2/sysfile.c b/fs/ocfs2/sysfile.c index bfe7190cdbf..902efb23b6a 100644 --- a/fs/ocfs2/sysfile.c +++ b/fs/ocfs2/sysfile.c @@ -44,11 +44,6 @@ static struct inode * _ocfs2_get_system_file_inode(struct ocfs2_super *osb, int type, u32 slot); -static inline int is_global_system_inode(int type); -static inline int is_in_system_inode_array(struct ocfs2_super *osb, - int type, - u32 slot); - #ifdef CONFIG_DEBUG_LOCK_ALLOC static struct lock_class_key ocfs2_sysfile_cluster_lock_key[NUM_SYSTEM_INODES]; #endif @@ -59,11 +54,52 @@ static inline int is_global_system_inode(int type) type <= OCFS2_LAST_GLOBAL_SYSTEM_INODE; } -static inline int is_in_system_inode_array(struct ocfs2_super *osb, - int type, - u32 slot) +static struct inode **get_local_system_inode(struct ocfs2_super *osb, + int type, + u32 slot) { - return slot == osb->slot_num || is_global_system_inode(type); + int index; + struct inode **local_system_inodes, **free = NULL; + + BUG_ON(slot == OCFS2_INVALID_SLOT); + BUG_ON(type < OCFS2_FIRST_LOCAL_SYSTEM_INODE || + type > OCFS2_LAST_LOCAL_SYSTEM_INODE); + + spin_lock(&osb->osb_lock); + local_system_inodes = osb->local_system_inodes; + spin_unlock(&osb->osb_lock); + + if (unlikely(!local_system_inodes)) { + local_system_inodes = kzalloc(sizeof(struct inode *) * + NUM_LOCAL_SYSTEM_INODES * + osb->max_slots, + GFP_NOFS); + if (!local_system_inodes) { + mlog_errno(-ENOMEM); + /* + * return NULL here so that ocfs2_get_sytem_file_inodes + * will try to create an inode and use it. We will try + * to initialize local_system_inodes next time. + */ + return NULL; + } + + spin_lock(&osb->osb_lock); + if (osb->local_system_inodes) { + /* Someone has initialized it for us. */ + free = local_system_inodes; + local_system_inodes = osb->local_system_inodes; + } else + osb->local_system_inodes = local_system_inodes; + spin_unlock(&osb->osb_lock); + if (unlikely(free)) + kfree(free); + } + + index = (slot * NUM_LOCAL_SYSTEM_INODES) + + (type - OCFS2_FIRST_LOCAL_SYSTEM_INODE); + + return &local_system_inodes[index]; } struct inode *ocfs2_get_system_file_inode(struct ocfs2_super *osb, @@ -74,8 +110,10 @@ struct inode *ocfs2_get_system_file_inode(struct ocfs2_super *osb, struct inode **arr = NULL; /* avoid the lookup if cached in local system file array */ - if (is_in_system_inode_array(osb, type, slot)) - arr = &(osb->system_inodes[type]); + if (is_global_system_inode(type)) { + arr = &(osb->global_system_inodes[type]); + } else + arr = get_local_system_inode(osb, type, slot); if (arr && ((inode = *arr) != NULL)) { /* get a ref in addition to the array ref */ diff --git a/fs/ocfs2/xattr.c b/fs/ocfs2/xattr.c index d03469f6180..67cd4391464 100644 --- a/fs/ocfs2/xattr.c +++ b/fs/ocfs2/xattr.c @@ -1286,13 +1286,11 @@ int ocfs2_xattr_get_nolock(struct inode *inode, xis.inode_bh = xbs.inode_bh = di_bh; di = (struct ocfs2_dinode *)di_bh->b_data; - down_read(&oi->ip_xattr_sem); ret = ocfs2_xattr_ibody_get(inode, name_index, name, buffer, buffer_size, &xis); if (ret == -ENODATA && di->i_xattr_loc) ret = ocfs2_xattr_block_get(inode, name_index, name, buffer, buffer_size, &xbs); - up_read(&oi->ip_xattr_sem); return ret; } @@ -1316,8 +1314,10 @@ static int ocfs2_xattr_get(struct inode *inode, mlog_errno(ret); return ret; } + down_read(&OCFS2_I(inode)->ip_xattr_sem); ret = ocfs2_xattr_get_nolock(inode, di_bh, name_index, name, buffer, buffer_size); + up_read(&OCFS2_I(inode)->ip_xattr_sem); ocfs2_inode_unlock(inode, 0); @@ -7081,7 +7081,7 @@ static int ocfs2_reflink_xattr_in_block(struct ocfs2_xattr_reflink *args, goto out; } - if (!(le16_to_cpu(xb->xb_flags) & OCFS2_XATTR_INDEXED)) + if (!indexed) ret = ocfs2_reflink_xattr_block(args, blk_bh, new_blk_bh); else ret = ocfs2_reflink_xattr_tree(args, blk_bh, new_blk_bh); |