diff options
Diffstat (limited to 'fs/xfs/xfs_log.c')
| -rw-r--r-- | fs/xfs/xfs_log.c | 2287 |
1 files changed, 1234 insertions, 1053 deletions
diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c index cee4ab9f8a9..292308dede6 100644 --- a/fs/xfs/xfs_log.c +++ b/fs/xfs/xfs_log.c @@ -17,162 +17,327 @@ */ #include "xfs.h" #include "xfs_fs.h" -#include "xfs_types.h" -#include "xfs_bit.h" -#include "xfs_log.h" -#include "xfs_inum.h" -#include "xfs_trans.h" +#include "xfs_shared.h" +#include "xfs_format.h" +#include "xfs_log_format.h" +#include "xfs_trans_resv.h" #include "xfs_sb.h" #include "xfs_ag.h" #include "xfs_mount.h" #include "xfs_error.h" +#include "xfs_trans.h" +#include "xfs_trans_priv.h" +#include "xfs_log.h" #include "xfs_log_priv.h" -#include "xfs_buf_item.h" -#include "xfs_bmap_btree.h" -#include "xfs_alloc_btree.h" -#include "xfs_ialloc_btree.h" #include "xfs_log_recover.h" -#include "xfs_trans_priv.h" -#include "xfs_dinode.h" #include "xfs_inode.h" -#include "xfs_rw.h" #include "xfs_trace.h" +#include "xfs_fsops.h" +#include "xfs_cksum.h" kmem_zone_t *xfs_log_ticket_zone; /* Local miscellaneous function prototypes */ -STATIC int xlog_commit_record(struct log *log, struct xlog_ticket *ticket, - xlog_in_core_t **, xfs_lsn_t *); -STATIC xlog_t * xlog_alloc_log(xfs_mount_t *mp, - xfs_buftarg_t *log_target, - xfs_daddr_t blk_offset, - int num_bblks); -STATIC int xlog_space_left(xlog_t *log, int cycle, int bytes); -STATIC int xlog_sync(xlog_t *log, xlog_in_core_t *iclog); -STATIC void xlog_dealloc_log(xlog_t *log); +STATIC int +xlog_commit_record( + struct xlog *log, + struct xlog_ticket *ticket, + struct xlog_in_core **iclog, + xfs_lsn_t *commitlsnp); + +STATIC struct xlog * +xlog_alloc_log( + struct xfs_mount *mp, + struct xfs_buftarg *log_target, + xfs_daddr_t blk_offset, + int num_bblks); +STATIC int +xlog_space_left( + struct xlog *log, + atomic64_t *head); +STATIC int +xlog_sync( + struct xlog *log, + struct xlog_in_core *iclog); +STATIC void +xlog_dealloc_log( + struct xlog *log); /* local state machine functions */ STATIC void xlog_state_done_syncing(xlog_in_core_t *iclog, int); -STATIC void xlog_state_do_callback(xlog_t *log,int aborted, xlog_in_core_t *iclog); -STATIC int xlog_state_get_iclog_space(xlog_t *log, - int len, - xlog_in_core_t **iclog, - xlog_ticket_t *ticket, - int *continued_write, - int *logoffsetp); -STATIC int xlog_state_release_iclog(xlog_t *log, - xlog_in_core_t *iclog); -STATIC void xlog_state_switch_iclogs(xlog_t *log, - xlog_in_core_t *iclog, - int eventual_size); -STATIC void xlog_state_want_sync(xlog_t *log, xlog_in_core_t *iclog); - -/* local functions to manipulate grant head */ -STATIC int xlog_grant_log_space(xlog_t *log, - xlog_ticket_t *xtic); -STATIC void xlog_grant_push_ail(xfs_mount_t *mp, - int need_bytes); -STATIC void xlog_regrant_reserve_log_space(xlog_t *log, - xlog_ticket_t *ticket); -STATIC int xlog_regrant_write_log_space(xlog_t *log, - xlog_ticket_t *ticket); -STATIC void xlog_ungrant_log_space(xlog_t *log, - xlog_ticket_t *ticket); +STATIC void +xlog_state_do_callback( + struct xlog *log, + int aborted, + struct xlog_in_core *iclog); +STATIC int +xlog_state_get_iclog_space( + struct xlog *log, + int len, + struct xlog_in_core **iclog, + struct xlog_ticket *ticket, + int *continued_write, + int *logoffsetp); +STATIC int +xlog_state_release_iclog( + struct xlog *log, + struct xlog_in_core *iclog); +STATIC void +xlog_state_switch_iclogs( + struct xlog *log, + struct xlog_in_core *iclog, + int eventual_size); +STATIC void +xlog_state_want_sync( + struct xlog *log, + struct xlog_in_core *iclog); + +STATIC void +xlog_grant_push_ail( + struct xlog *log, + int need_bytes); +STATIC void +xlog_regrant_reserve_log_space( + struct xlog *log, + struct xlog_ticket *ticket); +STATIC void +xlog_ungrant_log_space( + struct xlog *log, + struct xlog_ticket *ticket); #if defined(DEBUG) -STATIC void xlog_verify_dest_ptr(xlog_t *log, char *ptr); -STATIC void xlog_verify_grant_head(xlog_t *log, int equals); -STATIC void xlog_verify_iclog(xlog_t *log, xlog_in_core_t *iclog, - int count, boolean_t syncing); -STATIC void xlog_verify_tail_lsn(xlog_t *log, xlog_in_core_t *iclog, - xfs_lsn_t tail_lsn); +STATIC void +xlog_verify_dest_ptr( + struct xlog *log, + char *ptr); +STATIC void +xlog_verify_grant_tail( + struct xlog *log); +STATIC void +xlog_verify_iclog( + struct xlog *log, + struct xlog_in_core *iclog, + int count, + bool syncing); +STATIC void +xlog_verify_tail_lsn( + struct xlog *log, + struct xlog_in_core *iclog, + xfs_lsn_t tail_lsn); #else #define xlog_verify_dest_ptr(a,b) -#define xlog_verify_grant_head(a,b) +#define xlog_verify_grant_tail(a) #define xlog_verify_iclog(a,b,c,d) #define xlog_verify_tail_lsn(a,b,c) #endif -STATIC int xlog_iclogs_empty(xlog_t *log); - +STATIC int +xlog_iclogs_empty( + struct xlog *log); static void -xlog_ins_ticketq(struct xlog_ticket **qp, struct xlog_ticket *tic) +xlog_grant_sub_space( + struct xlog *log, + atomic64_t *head, + int bytes) { - if (*qp) { - tic->t_next = (*qp); - tic->t_prev = (*qp)->t_prev; - (*qp)->t_prev->t_next = tic; - (*qp)->t_prev = tic; - } else { - tic->t_prev = tic->t_next = tic; - *qp = tic; - } + int64_t head_val = atomic64_read(head); + int64_t new, old; - tic->t_flags |= XLOG_TIC_IN_Q; + do { + int cycle, space; + + xlog_crack_grant_head_val(head_val, &cycle, &space); + + space -= bytes; + if (space < 0) { + space += log->l_logsize; + cycle--; + } + + old = head_val; + new = xlog_assign_grant_head_val(cycle, space); + head_val = atomic64_cmpxchg(head, old, new); + } while (head_val != old); } static void -xlog_del_ticketq(struct xlog_ticket **qp, struct xlog_ticket *tic) +xlog_grant_add_space( + struct xlog *log, + atomic64_t *head, + int bytes) { - if (tic == tic->t_next) { - *qp = NULL; - } else { - *qp = tic->t_next; - tic->t_next->t_prev = tic->t_prev; - tic->t_prev->t_next = tic->t_next; - } + int64_t head_val = atomic64_read(head); + int64_t new, old; + + do { + int tmp; + int cycle, space; + + xlog_crack_grant_head_val(head_val, &cycle, &space); - tic->t_next = tic->t_prev = NULL; - tic->t_flags &= ~XLOG_TIC_IN_Q; + tmp = log->l_logsize - space; + if (tmp > bytes) + space += bytes; + else { + space = bytes - tmp; + cycle++; + } + + old = head_val; + new = xlog_assign_grant_head_val(cycle, space); + head_val = atomic64_cmpxchg(head, old, new); + } while (head_val != old); } -static void -xlog_grant_sub_space(struct log *log, int bytes) +STATIC void +xlog_grant_head_init( + struct xlog_grant_head *head) { - log->l_grant_write_bytes -= bytes; - if (log->l_grant_write_bytes < 0) { - log->l_grant_write_bytes += log->l_logsize; - log->l_grant_write_cycle--; - } + xlog_assign_grant_head(&head->grant, 1, 0); + INIT_LIST_HEAD(&head->waiters); + spin_lock_init(&head->lock); +} - log->l_grant_reserve_bytes -= bytes; - if ((log)->l_grant_reserve_bytes < 0) { - log->l_grant_reserve_bytes += log->l_logsize; - log->l_grant_reserve_cycle--; - } +STATIC void +xlog_grant_head_wake_all( + struct xlog_grant_head *head) +{ + struct xlog_ticket *tic; + spin_lock(&head->lock); + list_for_each_entry(tic, &head->waiters, t_queue) + wake_up_process(tic->t_task); + spin_unlock(&head->lock); } -static void -xlog_grant_add_space_write(struct log *log, int bytes) +static inline int +xlog_ticket_reservation( + struct xlog *log, + struct xlog_grant_head *head, + struct xlog_ticket *tic) { - int tmp = log->l_logsize - log->l_grant_write_bytes; - if (tmp > bytes) - log->l_grant_write_bytes += bytes; - else { - log->l_grant_write_cycle++; - log->l_grant_write_bytes = bytes - tmp; + if (head == &log->l_write_head) { + ASSERT(tic->t_flags & XLOG_TIC_PERM_RESERV); + return tic->t_unit_res; + } else { + if (tic->t_flags & XLOG_TIC_PERM_RESERV) + return tic->t_unit_res * tic->t_cnt; + else + return tic->t_unit_res; } } -static void -xlog_grant_add_space_reserve(struct log *log, int bytes) +STATIC bool +xlog_grant_head_wake( + struct xlog *log, + struct xlog_grant_head *head, + int *free_bytes) { - int tmp = log->l_logsize - log->l_grant_reserve_bytes; - if (tmp > bytes) - log->l_grant_reserve_bytes += bytes; - else { - log->l_grant_reserve_cycle++; - log->l_grant_reserve_bytes = bytes - tmp; + struct xlog_ticket *tic; + int need_bytes; + + list_for_each_entry(tic, &head->waiters, t_queue) { + need_bytes = xlog_ticket_reservation(log, head, tic); + if (*free_bytes < need_bytes) + return false; + + *free_bytes -= need_bytes; + trace_xfs_log_grant_wake_up(log, tic); + wake_up_process(tic->t_task); } + + return true; } -static inline void -xlog_grant_add_space(struct log *log, int bytes) +STATIC int +xlog_grant_head_wait( + struct xlog *log, + struct xlog_grant_head *head, + struct xlog_ticket *tic, + int need_bytes) __releases(&head->lock) + __acquires(&head->lock) { - xlog_grant_add_space_write(log, bytes); - xlog_grant_add_space_reserve(log, bytes); + list_add_tail(&tic->t_queue, &head->waiters); + + do { + if (XLOG_FORCED_SHUTDOWN(log)) + goto shutdown; + xlog_grant_push_ail(log, need_bytes); + + __set_current_state(TASK_UNINTERRUPTIBLE); + spin_unlock(&head->lock); + + XFS_STATS_INC(xs_sleep_logspace); + + trace_xfs_log_grant_sleep(log, tic); + schedule(); + trace_xfs_log_grant_wake(log, tic); + + spin_lock(&head->lock); + if (XLOG_FORCED_SHUTDOWN(log)) + goto shutdown; + } while (xlog_space_left(log, &head->grant) < need_bytes); + + list_del_init(&tic->t_queue); + return 0; +shutdown: + list_del_init(&tic->t_queue); + return XFS_ERROR(EIO); +} + +/* + * Atomically get the log space required for a log ticket. + * + * Once a ticket gets put onto head->waiters, it will only return after the + * needed reservation is satisfied. + * + * This function is structured so that it has a lock free fast path. This is + * necessary because every new transaction reservation will come through this + * path. Hence any lock will be globally hot if we take it unconditionally on + * every pass. + * + * As tickets are only ever moved on and off head->waiters under head->lock, we + * only need to take that lock if we are going to add the ticket to the queue + * and sleep. We can avoid taking the lock if the ticket was never added to + * head->waiters because the t_queue list head will be empty and we hold the + * only reference to it so it can safely be checked unlocked. + */ +STATIC int +xlog_grant_head_check( + struct xlog *log, + struct xlog_grant_head *head, + struct xlog_ticket *tic, + int *need_bytes) +{ + int free_bytes; + int error = 0; + + ASSERT(!(log->l_flags & XLOG_ACTIVE_RECOVERY)); + + /* + * If there are other waiters on the queue then give them a chance at + * logspace before us. Wake up the first waiters, if we do not wake + * up all the waiters then go to sleep waiting for more free space, + * otherwise try to get some space for this transaction. + */ + *need_bytes = xlog_ticket_reservation(log, head, tic); + free_bytes = xlog_space_left(log, &head->grant); + if (!list_empty_careful(&head->waiters)) { + spin_lock(&head->lock); + if (!xlog_grant_head_wake(log, head, &free_bytes) || + free_bytes < *need_bytes) { + error = xlog_grant_head_wait(log, head, tic, + *need_bytes); + } + spin_unlock(&head->lock); + } else if (free_bytes < *need_bytes) { + spin_lock(&head->lock); + error = xlog_grant_head_wait(log, head, tic, *need_bytes); + spin_unlock(&head->lock); + } + + return error; } static void @@ -200,6 +365,129 @@ xlog_tic_add_region(xlog_ticket_t *tic, uint len, uint type) } /* + * Replenish the byte reservation required by moving the grant write head. + */ +int +xfs_log_regrant( + struct xfs_mount *mp, + struct xlog_ticket *tic) +{ + struct xlog *log = mp->m_log; + int need_bytes; + int error = 0; + + if (XLOG_FORCED_SHUTDOWN(log)) + return XFS_ERROR(EIO); + + XFS_STATS_INC(xs_try_logspace); + + /* + * This is a new transaction on the ticket, so we need to change the + * transaction ID so that the next transaction has a different TID in + * the log. Just add one to the existing tid so that we can see chains + * of rolling transactions in the log easily. + */ + tic->t_tid++; + + xlog_grant_push_ail(log, tic->t_unit_res); + + tic->t_curr_res = tic->t_unit_res; + xlog_tic_reset_res(tic); + + if (tic->t_cnt > 0) + return 0; + + trace_xfs_log_regrant(log, tic); + + error = xlog_grant_head_check(log, &log->l_write_head, tic, + &need_bytes); + if (error) + goto out_error; + + xlog_grant_add_space(log, &log->l_write_head.grant, need_bytes); + trace_xfs_log_regrant_exit(log, tic); + xlog_verify_grant_tail(log); + return 0; + +out_error: + /* + * If we are failing, make sure the ticket doesn't have any current + * reservations. We don't want to add this back when the ticket/ + * transaction gets cancelled. + */ + tic->t_curr_res = 0; + tic->t_cnt = 0; /* ungrant will give back unit_res * t_cnt. */ + return error; +} + +/* + * Reserve log space and return a ticket corresponding the reservation. + * + * Each reservation is going to reserve extra space for a log record header. + * When writes happen to the on-disk log, we don't subtract the length of the + * log record header from any reservation. By wasting space in each + * reservation, we prevent over allocation problems. + */ +int +xfs_log_reserve( + struct xfs_mount *mp, + int unit_bytes, + int cnt, + struct xlog_ticket **ticp, + __uint8_t client, + bool permanent, + uint t_type) +{ + struct xlog *log = mp->m_log; + struct xlog_ticket *tic; + int need_bytes; + int error = 0; + + ASSERT(client == XFS_TRANSACTION || client == XFS_LOG); + + if (XLOG_FORCED_SHUTDOWN(log)) + return XFS_ERROR(EIO); + + XFS_STATS_INC(xs_try_logspace); + + ASSERT(*ticp == NULL); + tic = xlog_ticket_alloc(log, unit_bytes, cnt, client, permanent, + KM_SLEEP | KM_MAYFAIL); + if (!tic) + return XFS_ERROR(ENOMEM); + + tic->t_trans_type = t_type; + *ticp = tic; + + xlog_grant_push_ail(log, tic->t_cnt ? tic->t_unit_res * tic->t_cnt + : tic->t_unit_res); + + trace_xfs_log_reserve(log, tic); + + error = xlog_grant_head_check(log, &log->l_reserve_head, tic, + &need_bytes); + if (error) + goto out_error; + + xlog_grant_add_space(log, &log->l_reserve_head.grant, need_bytes); + xlog_grant_add_space(log, &log->l_write_head.grant, need_bytes); + trace_xfs_log_reserve_exit(log, tic); + xlog_verify_grant_tail(log); + return 0; + +out_error: + /* + * If we are failing, make sure the ticket doesn't have any current + * reservations. We don't want to add this back when the ticket/ + * transaction gets cancelled. + */ + tic->t_curr_res = 0; + tic->t_cnt = 0; /* ungrant will give back unit_res * t_cnt. */ + return error; +} + + +/* * NOTES: * * 1. currblock field gets updated at startup and after in-core logs @@ -227,7 +515,7 @@ xfs_log_done( struct xlog_in_core **iclog, uint flags) { - struct log *log = mp->m_log; + struct xlog *log = mp->m_log; xfs_lsn_t lsn = 0; if (XLOG_FORCED_SHUTDOWN(log) || @@ -309,77 +597,6 @@ xfs_log_release_iclog( } /* - * 1. Reserve an amount of on-disk log space and return a ticket corresponding - * to the reservation. - * 2. Potentially, push buffers at tail of log to disk. - * - * Each reservation is going to reserve extra space for a log record header. - * When writes happen to the on-disk log, we don't subtract the length of the - * log record header from any reservation. By wasting space in each - * reservation, we prevent over allocation problems. - */ -int -xfs_log_reserve( - struct xfs_mount *mp, - int unit_bytes, - int cnt, - struct xlog_ticket **ticket, - __uint8_t client, - uint flags, - uint t_type) -{ - struct log *log = mp->m_log; - struct xlog_ticket *internal_ticket; - int retval = 0; - - ASSERT(client == XFS_TRANSACTION || client == XFS_LOG); - - if (XLOG_FORCED_SHUTDOWN(log)) - return XFS_ERROR(EIO); - - XFS_STATS_INC(xs_try_logspace); - - - if (*ticket != NULL) { - ASSERT(flags & XFS_LOG_PERM_RESERV); - internal_ticket = *ticket; - - /* - * this is a new transaction on the ticket, so we need to - * change the transaction ID so that the next transaction has a - * different TID in the log. Just add one to the existing tid - * so that we can see chains of rolling transactions in the log - * easily. - */ - internal_ticket->t_tid++; - - trace_xfs_log_reserve(log, internal_ticket); - - xlog_grant_push_ail(mp, internal_ticket->t_unit_res); - retval = xlog_regrant_write_log_space(log, internal_ticket); - } else { - /* may sleep if need to allocate more tickets */ - internal_ticket = xlog_ticket_alloc(log, unit_bytes, cnt, - client, flags, - KM_SLEEP|KM_MAYFAIL); - if (!internal_ticket) - return XFS_ERROR(ENOMEM); - internal_ticket->t_trans_type = t_type; - *ticket = internal_ticket; - - trace_xfs_log_reserve(log, internal_ticket); - - xlog_grant_push_ail(mp, - (internal_ticket->t_unit_res * - internal_ticket->t_cnt)); - retval = xlog_grant_log_space(log, internal_ticket); - } - - return retval; -} /* xfs_log_reserve */ - - -/* * Mount a log filesystem * * mp - ubiquitous xfs mount point structure @@ -396,14 +613,16 @@ xfs_log_mount( xfs_daddr_t blk_offset, int num_bblks) { - int error; + int error = 0; + int min_logfsbs; - if (!(mp->m_flags & XFS_MOUNT_NORECOVERY)) - cmn_err(CE_NOTE, "XFS mounting filesystem %s", mp->m_fsname); - else { - cmn_err(CE_NOTE, - "!Mounting filesystem \"%s\" in no-recovery mode. Filesystem will be inconsistent.", - mp->m_fsname); + if (!(mp->m_flags & XFS_MOUNT_NORECOVERY)) { + xfs_notice(mp, "Mounting V%d Filesystem", + XFS_SB_VERSION_NUM(&mp->m_sb)); + } else { + xfs_notice(mp, +"Mounting V%d filesystem in no-recovery mode. Filesystem will be inconsistent.", + XFS_SB_VERSION_NUM(&mp->m_sb)); ASSERT(mp->m_flags & XFS_MOUNT_RDONLY); } @@ -414,11 +633,55 @@ xfs_log_mount( } /* + * Validate the given log space and drop a critical message via syslog + * if the log size is too small that would lead to some unexpected + * situations in transaction log space reservation stage. + * + * Note: we can't just reject the mount if the validation fails. This + * would mean that people would have to downgrade their kernel just to + * remedy the situation as there is no way to grow the log (short of + * black magic surgery with xfs_db). + * + * We can, however, reject mounts for CRC format filesystems, as the + * mkfs binary being used to make the filesystem should never create a + * filesystem with a log that is too small. + */ + min_logfsbs = xfs_log_calc_minimum_size(mp); + + if (mp->m_sb.sb_logblocks < min_logfsbs) { + xfs_warn(mp, + "Log size %d blocks too small, minimum size is %d blocks", + mp->m_sb.sb_logblocks, min_logfsbs); + error = EINVAL; + } else if (mp->m_sb.sb_logblocks > XFS_MAX_LOG_BLOCKS) { + xfs_warn(mp, + "Log size %d blocks too large, maximum size is %lld blocks", + mp->m_sb.sb_logblocks, XFS_MAX_LOG_BLOCKS); + error = EINVAL; + } else if (XFS_FSB_TO_B(mp, mp->m_sb.sb_logblocks) > XFS_MAX_LOG_BYTES) { + xfs_warn(mp, + "log size %lld bytes too large, maximum size is %lld bytes", + XFS_FSB_TO_B(mp, mp->m_sb.sb_logblocks), + XFS_MAX_LOG_BYTES); + error = EINVAL; + } + if (error) { + if (xfs_sb_version_hascrc(&mp->m_sb)) { + xfs_crit(mp, "AAIEEE! Log failed size checks. Abort!"); + ASSERT(0); + goto out_free_log; + } + xfs_crit(mp, +"Log size out of supported range. Continuing onwards, but if log hangs are\n" +"experienced then please report this message in the bug report."); + } + + /* * Initialize the AIL now we have a log. */ error = xfs_trans_ail_init(mp); if (error) { - cmn_err(CE_WARN, "XFS: AIL initialisation failed: error %d", error); + xfs_warn(mp, "AIL initialisation failed: error %d", error); goto out_free_log; } mp->m_log->l_ailp = mp->m_ail; @@ -438,7 +701,8 @@ xfs_log_mount( if (readonly) mp->m_flags |= XFS_MOUNT_RDONLY; if (error) { - cmn_err(CE_WARN, "XFS: log mount/recovery failed: error %d", error); + xfs_warn(mp, "log mount/recovery failed: error %d", + error); goto out_destroy_ail; } } @@ -464,25 +728,29 @@ out: } /* - * Finish the recovery of the file system. This is separate from - * the xfs_log_mount() call, because it depends on the code in - * xfs_mountfs() to read in the root and real-time bitmap inodes - * between calling xfs_log_mount() and here. + * Finish the recovery of the file system. This is separate from the + * xfs_log_mount() call, because it depends on the code in xfs_mountfs() to read + * in the root and real-time bitmap inodes between calling xfs_log_mount() and + * here. * - * mp - ubiquitous xfs mount point structure + * If we finish recovery successfully, start the background log work. If we are + * not doing recovery, then we have a RO filesystem and we don't need to start + * it. */ int xfs_log_mount_finish(xfs_mount_t *mp) { - int error; + int error = 0; - if (!(mp->m_flags & XFS_MOUNT_NORECOVERY)) + if (!(mp->m_flags & XFS_MOUNT_NORECOVERY)) { error = xlog_recover_finish(mp->m_log); - else { - error = 0; + if (!error) + xfs_log_work_queue(mp); + } else { ASSERT(mp->m_flags & XFS_MOUNT_RDONLY); } + return error; } @@ -498,14 +766,14 @@ xfs_log_mount_finish(xfs_mount_t *mp) * Unmount record used to have a string "Unmount filesystem--" in the * data section where the "Un" was really a magic number (XLOG_UNMOUNT_TYPE). * We just write the magic number now since that particular field isn't - * currently architecture converted and "nUmount" is a bit foo. + * currently architecture converted and "Unmount" is a bit foo. * As far as I know, there weren't any dependencies on the old behaviour. */ int xfs_log_unmount_write(xfs_mount_t *mp) { - xlog_t *log = mp->m_log; + struct xlog *log = mp->m_log; xlog_in_core_t *iclog; #ifdef DEBUG xlog_in_core_t *first_iclog; @@ -556,8 +824,9 @@ xfs_log_unmount_write(xfs_mount_t *mp) .lv_iovecp = ®, }; - /* remove inited flag */ + /* remove inited flag, and account for space used */ tic->t_flags = 0; + tic->t_curr_res -= sizeof(magic); error = xlog_write(log, &vec, tic, &lsn, NULL, XLOG_UNMOUNT_TRANS); /* @@ -567,10 +836,8 @@ xfs_log_unmount_write(xfs_mount_t *mp) */ } - if (error) { - xfs_fs_cmn_err(CE_ALERT, mp, - "xfs_log_unmount: unmount record failed"); - } + if (error) + xfs_alert(mp, "%s: unmount record failed", __func__); spin_lock(&log->l_icloglock); @@ -584,8 +851,8 @@ xfs_log_unmount_write(xfs_mount_t *mp) if (!(iclog->ic_state == XLOG_STATE_ACTIVE || iclog->ic_state == XLOG_STATE_DIRTY)) { if (!XLOG_FORCED_SHUTDOWN(log)) { - sv_wait(&iclog->ic_force_wait, PMEM, - &log->l_icloglock, s); + xlog_wait(&iclog->ic_force_wait, + &log->l_icloglock); } else { spin_unlock(&log->l_icloglock); } @@ -625,8 +892,8 @@ xfs_log_unmount_write(xfs_mount_t *mp) || iclog->ic_state == XLOG_STATE_DIRTY || iclog->ic_state == XLOG_STATE_IOERROR) ) { - sv_wait(&iclog->ic_force_wait, PMEM, - &log->l_icloglock, s); + xlog_wait(&iclog->ic_force_wait, + &log->l_icloglock); } else { spin_unlock(&log->l_icloglock); } @@ -636,14 +903,49 @@ xfs_log_unmount_write(xfs_mount_t *mp) } /* xfs_log_unmount_write */ /* - * Deallocate log structures for unmount/relocation. + * Empty the log for unmount/freeze. * - * We need to stop the aild from running before we destroy - * and deallocate the log as the aild references the log. + * To do this, we first need to shut down the background log work so it is not + * trying to cover the log as we clean up. We then need to unpin all objects in + * the log so we can then flush them out. Once they have completed their IO and + * run the callbacks removing themselves from the AIL, we can write the unmount + * record. */ void -xfs_log_unmount(xfs_mount_t *mp) +xfs_log_quiesce( + struct xfs_mount *mp) { + cancel_delayed_work_sync(&mp->m_log->l_work); + xfs_log_force(mp, XFS_LOG_SYNC); + + /* + * The superblock buffer is uncached and while xfs_ail_push_all_sync() + * will push it, xfs_wait_buftarg() will not wait for it. Further, + * xfs_buf_iowait() cannot be used because it was pushed with the + * XBF_ASYNC flag set, so we need to use a lock/unlock pair to wait for + * the IO to complete. + */ + xfs_ail_push_all_sync(mp->m_ail); + xfs_wait_buftarg(mp->m_ddev_targp); + xfs_buf_lock(mp->m_sb_bp); + xfs_buf_unlock(mp->m_sb_bp); + + xfs_log_unmount_write(mp); +} + +/* + * Shut down and release the AIL and Log. + * + * During unmount, we need to ensure we flush all the dirty metadata objects + * from the AIL so that the log is empty before we write the unmount record to + * the log. Once this is done, we can tear down the AIL and the log. + */ +void +xfs_log_unmount( + struct xfs_mount *mp) +{ + xfs_log_quiesce(mp); + xfs_trans_ail_destroy(mp); xlog_dealloc_log(mp->m_log); } @@ -653,7 +955,7 @@ xfs_log_item_init( struct xfs_mount *mp, struct xfs_log_item *item, int type, - struct xfs_item_ops *ops) + const struct xfs_item_ops *ops) { item->li_mountp = mp; item->li_ailp = mp->m_ail; @@ -666,129 +968,66 @@ xfs_log_item_init( } /* - * Write region vectors to log. The write happens using the space reservation - * of the ticket (tic). It is not a requirement that all writes for a given - * transaction occur with one call to xfs_log_write(). However, it is important - * to note that the transaction reservation code makes an assumption about the - * number of log headers a transaction requires that may be violated if you - * don't pass all the transaction vectors in one call.... + * Wake up processes waiting for log space after we have moved the log tail. */ -int -xfs_log_write( - struct xfs_mount *mp, - struct xfs_log_iovec reg[], - int nentries, - struct xlog_ticket *tic, - xfs_lsn_t *start_lsn) -{ - struct log *log = mp->m_log; - int error; - struct xfs_log_vec vec = { - .lv_niovecs = nentries, - .lv_iovecp = reg, - }; - - if (XLOG_FORCED_SHUTDOWN(log)) - return XFS_ERROR(EIO); - - error = xlog_write(log, &vec, tic, start_lsn, NULL, 0); - if (error) - xfs_force_shutdown(mp, SHUTDOWN_LOG_IO_ERROR); - return error; -} - void -xfs_log_move_tail(xfs_mount_t *mp, - xfs_lsn_t tail_lsn) +xfs_log_space_wake( + struct xfs_mount *mp) { - xlog_ticket_t *tic; - xlog_t *log = mp->m_log; - int need_bytes, free_bytes, cycle, bytes; + struct xlog *log = mp->m_log; + int free_bytes; if (XLOG_FORCED_SHUTDOWN(log)) return; - if (tail_lsn == 0) { - /* needed since sync_lsn is 64 bits */ - spin_lock(&log->l_icloglock); - tail_lsn = log->l_last_sync_lsn; - spin_unlock(&log->l_icloglock); - } - - spin_lock(&log->l_grant_lock); + if (!list_empty_careful(&log->l_write_head.waiters)) { + ASSERT(!(log->l_flags & XLOG_ACTIVE_RECOVERY)); - /* Also an invalid lsn. 1 implies that we aren't passing in a valid - * tail_lsn. - */ - if (tail_lsn != 1) { - log->l_tail_lsn = tail_lsn; + spin_lock(&log->l_write_head.lock); + free_bytes = xlog_space_left(log, &log->l_write_head.grant); + xlog_grant_head_wake(log, &log->l_write_head, &free_bytes); + spin_unlock(&log->l_write_head.lock); } - if ((tic = log->l_write_headq)) { -#ifdef DEBUG - if (log->l_flags & XLOG_ACTIVE_RECOVERY) - panic("Recovery problem"); -#endif - cycle = log->l_grant_write_cycle; - bytes = log->l_grant_write_bytes; - free_bytes = xlog_space_left(log, cycle, bytes); - do { - ASSERT(tic->t_flags & XLOG_TIC_PERM_RESERV); + if (!list_empty_careful(&log->l_reserve_head.waiters)) { + ASSERT(!(log->l_flags & XLOG_ACTIVE_RECOVERY)); - if (free_bytes < tic->t_unit_res && tail_lsn != 1) - break; - tail_lsn = 0; - free_bytes -= tic->t_unit_res; - sv_signal(&tic->t_wait); - tic = tic->t_next; - } while (tic != log->l_write_headq); - } - if ((tic = log->l_reserve_headq)) { -#ifdef DEBUG - if (log->l_flags & XLOG_ACTIVE_RECOVERY) - panic("Recovery problem"); -#endif - cycle = log->l_grant_reserve_cycle; - bytes = log->l_grant_reserve_bytes; - free_bytes = xlog_space_left(log, cycle, bytes); - do { - if (tic->t_flags & XLOG_TIC_PERM_RESERV) - need_bytes = tic->t_unit_res*tic->t_cnt; - else - need_bytes = tic->t_unit_res; - if (free_bytes < need_bytes && tail_lsn != 1) - break; - tail_lsn = 0; - free_bytes -= need_bytes; - sv_signal(&tic->t_wait); - tic = tic->t_next; - } while (tic != log->l_reserve_headq); + spin_lock(&log->l_reserve_head.lock); + free_bytes = xlog_space_left(log, &log->l_reserve_head.grant); + xlog_grant_head_wake(log, &log->l_reserve_head, &free_bytes); + spin_unlock(&log->l_reserve_head.lock); } - spin_unlock(&log->l_grant_lock); -} /* xfs_log_move_tail */ +} /* - * Determine if we have a transaction that has gone to disk - * that needs to be covered. To begin the transition to the idle state - * firstly the log needs to be idle (no AIL and nothing in the iclogs). - * If we are then in a state where covering is needed, the caller is informed - * that dummy transactions are required to move the log into the idle state. + * Determine if we have a transaction that has gone to disk that needs to be + * covered. To begin the transition to the idle state firstly the log needs to + * be idle. That means the CIL, the AIL and the iclogs needs to be empty before + * we start attempting to cover the log. * - * Because this is called as part of the sync process, we should also indicate - * that dummy transactions should be issued in anything but the covered or - * idle states. This ensures that the log tail is accurately reflected in - * the log at the end of the sync, hence if a crash occurrs avoids replay - * of transactions where the metadata is already on disk. + * Only if we are then in a state where covering is needed, the caller is + * informed that dummy transactions are required to move the log into the idle + * state. + * + * If there are any items in the AIl or CIL, then we do not want to attempt to + * cover the log as we may be in a situation where there isn't log space + * available to run a dummy transaction and this can lead to deadlocks when the + * tail of the log is pinned by an item that is modified in the CIL. Hence + * there's no point in running a dummy transaction at this point because we + * can't start trying to idle the log until both the CIL and AIL are empty. */ int xfs_log_need_covered(xfs_mount_t *mp) { + struct xlog *log = mp->m_log; int needed = 0; - xlog_t *log = mp->m_log; if (!xfs_fs_writable(mp)) return 0; + if (!xlog_cil_empty(log)) + return 0; + spin_lock(&log->l_icloglock); switch (log->l_covered_state) { case XLOG_STATE_COVER_DONE: @@ -797,14 +1036,17 @@ xfs_log_need_covered(xfs_mount_t *mp) break; case XLOG_STATE_COVER_NEED: case XLOG_STATE_COVER_NEED2: - if (!xfs_trans_ail_tail(log->l_ailp) && - xlog_iclogs_empty(log)) { - if (log->l_covered_state == XLOG_STATE_COVER_NEED) - log->l_covered_state = XLOG_STATE_COVER_DONE; - else - log->l_covered_state = XLOG_STATE_COVER_DONE2; - } - /* FALLTHRU */ + if (xfs_ail_min_lsn(log->l_ailp)) + break; + if (!xlog_iclogs_empty(log)) + break; + + needed = 1; + if (log->l_covered_state == XLOG_STATE_COVER_NEED) + log->l_covered_state = XLOG_STATE_COVER_DONE; + else + log->l_covered_state = XLOG_STATE_COVER_DONE2; + break; default: needed = 1; break; @@ -813,41 +1055,46 @@ xfs_log_need_covered(xfs_mount_t *mp) return needed; } -/****************************************************************************** - * - * local routines - * - ****************************************************************************** - */ - -/* xfs_trans_tail_ail returns 0 when there is nothing in the list. - * The log manager must keep track of the last LR which was committed - * to disk. The lsn of this LR will become the new tail_lsn whenever - * xfs_trans_tail_ail returns 0. If we don't do this, we run into - * the situation where stuff could be written into the log but nothing - * was ever in the AIL when asked. Eventually, we panic since the - * tail hits the head. - * +/* * We may be holding the log iclog lock upon entering this routine. */ xfs_lsn_t -xlog_assign_tail_lsn(xfs_mount_t *mp) +xlog_assign_tail_lsn_locked( + struct xfs_mount *mp) { - xfs_lsn_t tail_lsn; - xlog_t *log = mp->m_log; + struct xlog *log = mp->m_log; + struct xfs_log_item *lip; + xfs_lsn_t tail_lsn; - tail_lsn = xfs_trans_ail_tail(mp->m_ail); - spin_lock(&log->l_grant_lock); - if (tail_lsn != 0) { - log->l_tail_lsn = tail_lsn; - } else { - tail_lsn = log->l_tail_lsn = log->l_last_sync_lsn; - } - spin_unlock(&log->l_grant_lock); + assert_spin_locked(&mp->m_ail->xa_lock); + /* + * To make sure we always have a valid LSN for the log tail we keep + * track of the last LSN which was committed in log->l_last_sync_lsn, + * and use that when the AIL was empty. + */ + lip = xfs_ail_min(mp->m_ail); + if (lip) + tail_lsn = lip->li_lsn; + else + tail_lsn = atomic64_read(&log->l_last_sync_lsn); + trace_xfs_log_assign_tail_lsn(log, tail_lsn); + atomic64_set(&log->l_tail_lsn, tail_lsn); return tail_lsn; -} /* xlog_assign_tail_lsn */ +} +xfs_lsn_t +xlog_assign_tail_lsn( + struct xfs_mount *mp) +{ + xfs_lsn_t tail_lsn; + + spin_lock(&mp->m_ail->xa_lock); + tail_lsn = xlog_assign_tail_lsn_locked(mp); + spin_unlock(&mp->m_ail->xa_lock); + + return tail_lsn; +} /* * Return the space in the log between the tail and the head. The head @@ -864,37 +1111,42 @@ xlog_assign_tail_lsn(xfs_mount_t *mp) * result is that we return the size of the log as the amount of space left. */ STATIC int -xlog_space_left(xlog_t *log, int cycle, int bytes) -{ - int free_bytes; - int tail_bytes; - int tail_cycle; - - tail_bytes = BBTOB(BLOCK_LSN(log->l_tail_lsn)); - tail_cycle = CYCLE_LSN(log->l_tail_lsn); - if ((tail_cycle == cycle) && (bytes >= tail_bytes)) { - free_bytes = log->l_logsize - (bytes - tail_bytes); - } else if ((tail_cycle + 1) < cycle) { +xlog_space_left( + struct xlog *log, + atomic64_t *head) +{ + int free_bytes; + int tail_bytes; + int tail_cycle; + int head_cycle; + int head_bytes; + + xlog_crack_grant_head(head, &head_cycle, &head_bytes); + xlog_crack_atomic_lsn(&log->l_tail_lsn, &tail_cycle, &tail_bytes); + tail_bytes = BBTOB(tail_bytes); + if (tail_cycle == head_cycle && head_bytes >= tail_bytes) + free_bytes = log->l_logsize - (head_bytes - tail_bytes); + else if (tail_cycle + 1 < head_cycle) return 0; - } else if (tail_cycle < cycle) { - ASSERT(tail_cycle == (cycle - 1)); - free_bytes = tail_bytes - bytes; + else if (tail_cycle < head_cycle) { + ASSERT(tail_cycle == (head_cycle - 1)); + free_bytes = tail_bytes - head_bytes; } else { /* * The reservation head is behind the tail. * In this case we just want to return the size of the * log as the amount of space left. */ - xfs_fs_cmn_err(CE_ALERT, log->l_mp, + xfs_alert(log->l_mp, "xlog_space_left: head behind tail\n" " tail_cycle = %d, tail_bytes = %d\n" " GH cycle = %d, GH bytes = %d", - tail_cycle, tail_bytes, cycle, bytes); + tail_cycle, tail_bytes, head_cycle, head_bytes); ASSERT(0); free_bytes = log->l_logsize; } return free_bytes; -} /* xlog_space_left */ +} /* @@ -906,23 +1158,17 @@ xlog_space_left(xlog_t *log, int cycle, int bytes) void xlog_iodone(xfs_buf_t *bp) { - xlog_in_core_t *iclog; - xlog_t *l; - int aborted; - - iclog = XFS_BUF_FSPRIVATE(bp, xlog_in_core_t *); - ASSERT(XFS_BUF_FSPRIVATE2(bp, unsigned long) == (unsigned long) 2); - XFS_BUF_SET_FSPRIVATE2(bp, (unsigned long)1); - aborted = 0; - l = iclog->ic_log; + struct xlog_in_core *iclog = bp->b_fspriv; + struct xlog *l = iclog->ic_log; + int aborted = 0; /* * Race to shutdown the filesystem if we see an error. */ - if (XFS_TEST_ERROR((XFS_BUF_GETERROR(bp)), l->l_mp, + if (XFS_TEST_ERROR(bp->b_error, l->l_mp, XFS_ERRTAG_IODONE_IOERR, XFS_RANDOM_IODONE_IOERR)) { - xfs_ioerror_alert("xlog_iodone", l->l_mp, bp, XFS_BUF_ADDR(bp)); - XFS_BUF_STALE(bp); + xfs_buf_ioerror_alert(bp, __func__); + xfs_buf_stale(bp); xfs_force_shutdown(l->l_mp, SHUTDOWN_LOG_IO_ERROR); /* * This flag will be propagated to the trans-committed @@ -937,13 +1183,15 @@ xlog_iodone(xfs_buf_t *bp) /* log I/O is always issued ASYNC */ ASSERT(XFS_BUF_ISASYNC(bp)); xlog_state_done_syncing(iclog, aborted); + /* - * do not reference the buffer (bp) here as we could race - * with it being freed after writing the unmount record to the - * log. + * drop the buffer lock now that we are done. Nothing references + * the buffer after this, so an unmount waiting on this lock can now + * tear it down safely. As such, it is unsafe to reference the buffer + * (bp) after the unlock as we could race with it being freed. */ - -} /* xlog_iodone */ + xfs_buf_unlock(bp); +} /* * Return size of each in-core log record buffer. @@ -955,8 +1203,9 @@ xlog_iodone(xfs_buf_t *bp) */ STATIC void -xlog_get_iclog_buffer_size(xfs_mount_t *mp, - xlog_t *log) +xlog_get_iclog_buffer_size( + struct xfs_mount *mp, + struct xlog *log) { int size; int xhdrs; @@ -1012,18 +1261,53 @@ done: } /* xlog_get_iclog_buffer_size */ +void +xfs_log_work_queue( + struct xfs_mount *mp) +{ + queue_delayed_work(mp->m_log_workqueue, &mp->m_log->l_work, + msecs_to_jiffies(xfs_syncd_centisecs * 10)); +} + +/* + * Every sync period we need to unpin all items in the AIL and push them to + * disk. If there is nothing dirty, then we might need to cover the log to + * indicate that the filesystem is idle. + */ +void +xfs_log_worker( + struct work_struct *work) +{ + struct xlog *log = container_of(to_delayed_work(work), + struct xlog, l_work); + struct xfs_mount *mp = log->l_mp; + + /* dgc: errors ignored - not fatal and nowhere to report them */ + if (xfs_log_need_covered(mp)) + xfs_fs_log_dummy(mp); + else + xfs_log_force(mp, 0); + + /* start pushing all the metadata that is currently dirty */ + xfs_ail_push_all(mp->m_ail); + + /* queue us up again */ + xfs_log_work_queue(mp); +} + /* * This routine initializes some of the log structure for a given mount point. * Its primary purpose is to fill in enough, so recovery can occur. However, * some other stuff may be filled in too. */ -STATIC xlog_t * -xlog_alloc_log(xfs_mount_t *mp, - xfs_buftarg_t *log_target, - xfs_daddr_t blk_offset, - int num_bblks) +STATIC struct xlog * +xlog_alloc_log( + struct xfs_mount *mp, + struct xfs_buftarg *log_target, + xfs_daddr_t blk_offset, + int num_bblks) { - xlog_t *log; + struct xlog *log; xlog_rec_header_t *head; xlog_in_core_t **iclogp; xlog_in_core_t *iclog, *prev_iclog=NULL; @@ -1032,9 +1316,9 @@ xlog_alloc_log(xfs_mount_t *mp, int error = ENOMEM; uint log2_size = 0; - log = kmem_zalloc(sizeof(xlog_t), KM_MAYFAIL); + log = kmem_zalloc(sizeof(struct xlog), KM_MAYFAIL); if (!log) { - xlog_warn("XFS: Log allocation failed: No memory!"); + xfs_warn(mp, "Log allocation failed: No memory!"); goto out; } @@ -1045,37 +1329,39 @@ xlog_alloc_log(xfs_mount_t *mp, log->l_logBBsize = num_bblks; log->l_covered_state = XLOG_STATE_COVER_IDLE; log->l_flags |= XLOG_ACTIVE_RECOVERY; + INIT_DELAYED_WORK(&log->l_work, xfs_log_worker); log->l_prev_block = -1; - log->l_tail_lsn = xlog_assign_lsn(1, 0); /* log->l_tail_lsn = 0x100000000LL; cycle = 1; current block = 0 */ - log->l_last_sync_lsn = log->l_tail_lsn; + xlog_assign_atomic_lsn(&log->l_tail_lsn, 1, 0); + xlog_assign_atomic_lsn(&log->l_last_sync_lsn, 1, 0); log->l_curr_cycle = 1; /* 0 is bad since this is initial value */ - log->l_grant_reserve_cycle = 1; - log->l_grant_write_cycle = 1; + + xlog_grant_head_init(&log->l_reserve_head); + xlog_grant_head_init(&log->l_write_head); error = EFSCORRUPTED; if (xfs_sb_version_hassector(&mp->m_sb)) { log2_size = mp->m_sb.sb_logsectlog; if (log2_size < BBSHIFT) { - xlog_warn("XFS: Log sector size too small " - "(0x%x < 0x%x)", log2_size, BBSHIFT); + xfs_warn(mp, "Log sector size too small (0x%x < 0x%x)", + log2_size, BBSHIFT); goto out_free_log; } log2_size -= BBSHIFT; if (log2_size > mp->m_sectbb_log) { - xlog_warn("XFS: Log sector size too large " - "(0x%x > 0x%x)", log2_size, mp->m_sectbb_log); + xfs_warn(mp, "Log sector size too large (0x%x > 0x%x)", + log2_size, mp->m_sectbb_log); goto out_free_log; } /* for larger sector sizes, must have v2 or external log */ if (log2_size && log->l_logBBstart > 0 && !xfs_sb_version_haslogv2(&mp->m_sb)) { - - xlog_warn("XFS: log sector size (0x%x) invalid " - "for configuration.", log2_size); + xfs_warn(mp, + "log sector size (0x%x) invalid for configuration.", + log2_size); goto out_free_log; } } @@ -1084,21 +1370,23 @@ xlog_alloc_log(xfs_mount_t *mp, xlog_get_iclog_buffer_size(mp, log); error = ENOMEM; - bp = xfs_buf_get_empty(log->l_iclog_size, mp->m_logdev_targp); + bp = xfs_buf_alloc(mp->m_logdev_targp, 0, BTOBB(log->l_iclog_size), 0); if (!bp) goto out_free_log; - XFS_BUF_SET_IODONE_FUNC(bp, xlog_iodone); - XFS_BUF_SET_FSPRIVATE2(bp, (unsigned long)1); - ASSERT(XFS_BUF_ISBUSY(bp)); - ASSERT(XFS_BUF_VALUSEMA(bp) <= 0); + + /* + * The iclogbuf buffer locks are held over IO but we are not going to do + * IO yet. Hence unlock the buffer so that the log IO path can grab it + * when appropriately. + */ + ASSERT(xfs_buf_islocked(bp)); + xfs_buf_unlock(bp); + + bp->b_iodone = xlog_iodone; log->l_xbuf = bp; spin_lock_init(&log->l_icloglock); - spin_lock_init(&log->l_grant_lock); - sv_init(&log->l_flush_wait, 0, "flush_wait"); - - /* log record size must be multiple of BBSIZE; see xlog_rec_header_t */ - ASSERT((XFS_BUF_SIZE(bp) & BBMASK) == 0); + init_waitqueue_head(&log->l_flush_wait); iclogp = &log->l_iclog; /* @@ -1119,13 +1407,14 @@ xlog_alloc_log(xfs_mount_t *mp, prev_iclog = iclog; bp = xfs_buf_get_uncached(mp->m_logdev_targp, - log->l_iclog_size, 0); + BTOBB(log->l_iclog_size), 0); if (!bp) goto out_free_iclog; - if (!XFS_BUF_CPSEMA(bp)) - ASSERT(0); - XFS_BUF_SET_IODONE_FUNC(bp, xlog_iodone); - XFS_BUF_SET_FSPRIVATE2(bp, (unsigned long)1); + + ASSERT(xfs_buf_islocked(bp)); + xfs_buf_unlock(bp); + + bp->b_iodone = xlog_iodone; iclog->ic_bp = bp; iclog->ic_data = bp->b_addr; #ifdef DEBUG @@ -1141,7 +1430,7 @@ xlog_alloc_log(xfs_mount_t *mp, head->h_fmt = cpu_to_be32(XLOG_FMT); memcpy(&head->h_fs_uuid, &mp->m_sb.sb_uuid, sizeof(uuid_t)); - iclog->ic_size = XFS_BUF_SIZE(bp) - log->l_iclog_hsize; + iclog->ic_size = BBTOB(bp->b_length) - log->l_iclog_hsize; iclog->ic_state = XLOG_STATE_ACTIVE; iclog->ic_log = log; atomic_set(&iclog->ic_refcnt, 0); @@ -1149,10 +1438,8 @@ xlog_alloc_log(xfs_mount_t *mp, iclog->ic_callback_tail = &(iclog->ic_callback); iclog->ic_datap = (char *)iclog->ic_data + log->l_iclog_hsize; - ASSERT(XFS_BUF_ISBUSY(iclog->ic_bp)); - ASSERT(XFS_BUF_VALUSEMA(iclog->ic_bp) <= 0); - sv_init(&iclog->ic_force_wait, SV_DEFAULT, "iclog-force"); - sv_init(&iclog->ic_write_wait, SV_DEFAULT, "iclog-write"); + init_waitqueue_head(&iclog->ic_force_wait); + init_waitqueue_head(&iclog->ic_write_wait); iclogp = &iclog->ic_next; } @@ -1167,15 +1454,11 @@ xlog_alloc_log(xfs_mount_t *mp, out_free_iclog: for (iclog = log->l_iclog; iclog; iclog = prev_iclog) { prev_iclog = iclog->ic_next; - if (iclog->ic_bp) { - sv_destroy(&iclog->ic_force_wait); - sv_destroy(&iclog->ic_write_wait); + if (iclog->ic_bp) xfs_buf_free(iclog->ic_bp); - } kmem_free(iclog); } spinlock_destroy(&log->l_icloglock); - spinlock_destroy(&log->l_grant_lock); xfs_buf_free(log->l_xbuf); out_free_log: kmem_free(log); @@ -1190,7 +1473,7 @@ out: */ STATIC int xlog_commit_record( - struct log *log, + struct xlog *log, struct xlog_ticket *ticket, struct xlog_in_core **iclog, xfs_lsn_t *commitlsnp) @@ -1223,61 +1506,138 @@ xlog_commit_record( * water mark. In this manner, we would be creating a low water mark. */ STATIC void -xlog_grant_push_ail(xfs_mount_t *mp, - int need_bytes) -{ - xlog_t *log = mp->m_log; /* pointer to the log */ - xfs_lsn_t tail_lsn; /* lsn of the log tail */ - xfs_lsn_t threshold_lsn = 0; /* lsn we'd like to be at */ - int free_blocks; /* free blocks left to write to */ - int free_bytes; /* free bytes left to write to */ - int threshold_block; /* block in lsn we'd like to be at */ - int threshold_cycle; /* lsn cycle we'd like to be at */ - int free_threshold; - - ASSERT(BTOBB(need_bytes) < log->l_logBBsize); - - spin_lock(&log->l_grant_lock); - free_bytes = xlog_space_left(log, - log->l_grant_reserve_cycle, - log->l_grant_reserve_bytes); - tail_lsn = log->l_tail_lsn; - free_blocks = BTOBBT(free_bytes); - - /* - * Set the threshold for the minimum number of free blocks in the - * log to the maximum of what the caller needs, one quarter of the - * log, and 256 blocks. - */ - free_threshold = BTOBB(need_bytes); - free_threshold = MAX(free_threshold, (log->l_logBBsize >> 2)); - free_threshold = MAX(free_threshold, 256); - if (free_blocks < free_threshold) { - threshold_block = BLOCK_LSN(tail_lsn) + free_threshold; - threshold_cycle = CYCLE_LSN(tail_lsn); +xlog_grant_push_ail( + struct xlog *log, + int need_bytes) +{ + xfs_lsn_t threshold_lsn = 0; + xfs_lsn_t last_sync_lsn; + int free_blocks; + int free_bytes; + int threshold_block; + int threshold_cycle; + int free_threshold; + + ASSERT(BTOBB(need_bytes) < log->l_logBBsize); + + free_bytes = xlog_space_left(log, &log->l_reserve_head.grant); + free_blocks = BTOBBT(free_bytes); + + /* + * Set the threshold for the minimum number of free blocks in the + * log to the maximum of what the caller needs, one quarter of the + * log, and 256 blocks. + */ + free_threshold = BTOBB(need_bytes); + free_threshold = MAX(free_threshold, (log->l_logBBsize >> 2)); + free_threshold = MAX(free_threshold, 256); + if (free_blocks >= free_threshold) + return; + + xlog_crack_atomic_lsn(&log->l_tail_lsn, &threshold_cycle, + &threshold_block); + threshold_block += free_threshold; if (threshold_block >= log->l_logBBsize) { - threshold_block -= log->l_logBBsize; - threshold_cycle += 1; + threshold_block -= log->l_logBBsize; + threshold_cycle += 1; } - threshold_lsn = xlog_assign_lsn(threshold_cycle, threshold_block); + threshold_lsn = xlog_assign_lsn(threshold_cycle, + threshold_block); + /* + * Don't pass in an lsn greater than the lsn of the last + * log record known to be on disk. Use a snapshot of the last sync lsn + * so that it doesn't change between the compare and the set. + */ + last_sync_lsn = atomic64_read(&log->l_last_sync_lsn); + if (XFS_LSN_CMP(threshold_lsn, last_sync_lsn) > 0) + threshold_lsn = last_sync_lsn; - /* Don't pass in an lsn greater than the lsn of the last - * log record known to be on disk. + /* + * Get the transaction layer to kick the dirty buffers out to + * disk asynchronously. No point in trying to do this if + * the filesystem is shutting down. */ - if (XFS_LSN_CMP(threshold_lsn, log->l_last_sync_lsn) > 0) - threshold_lsn = log->l_last_sync_lsn; - } - spin_unlock(&log->l_grant_lock); - - /* - * Get the transaction layer to kick the dirty buffers out to - * disk asynchronously. No point in trying to do this if - * the filesystem is shutting down. - */ - if (threshold_lsn && - !XLOG_FORCED_SHUTDOWN(log)) - xfs_trans_ail_push(log->l_ailp, threshold_lsn); -} /* xlog_grant_push_ail */ + if (!XLOG_FORCED_SHUTDOWN(log)) + xfs_ail_push(log->l_ailp, threshold_lsn); +} + +/* + * Stamp cycle number in every block + */ +STATIC void +xlog_pack_data( + struct xlog *log, + struct xlog_in_core *iclog, + int roundoff) +{ + int i, j, k; + int size = iclog->ic_offset + roundoff; + __be32 cycle_lsn; + xfs_caddr_t dp; + + cycle_lsn = CYCLE_LSN_DISK(iclog->ic_header.h_lsn); + + dp = iclog->ic_datap; + for (i = 0; i < BTOBB(size); i++) { + if (i >= (XLOG_HEADER_CYCLE_SIZE / BBSIZE)) + break; + iclog->ic_header.h_cycle_data[i] = *(__be32 *)dp; + *(__be32 *)dp = cycle_lsn; + dp += BBSIZE; + } + + if (xfs_sb_version_haslogv2(&log->l_mp->m_sb)) { + xlog_in_core_2_t *xhdr = iclog->ic_data; + + for ( ; i < BTOBB(size); i++) { + j = i / (XLOG_HEADER_CYCLE_SIZE / BBSIZE); + k = i % (XLOG_HEADER_CYCLE_SIZE / BBSIZE); + xhdr[j].hic_xheader.xh_cycle_data[k] = *(__be32 *)dp; + *(__be32 *)dp = cycle_lsn; + dp += BBSIZE; + } + + for (i = 1; i < log->l_iclog_heads; i++) + xhdr[i].hic_xheader.xh_cycle = cycle_lsn; + } +} + +/* + * Calculate the checksum for a log buffer. + * + * This is a little more complicated than it should be because the various + * headers and the actual data are non-contiguous. + */ +__le32 +xlog_cksum( + struct xlog *log, + struct xlog_rec_header *rhead, + char *dp, + int size) +{ + __uint32_t crc; + + /* first generate the crc for the record header ... */ + crc = xfs_start_cksum((char *)rhead, + sizeof(struct xlog_rec_header), + offsetof(struct xlog_rec_header, h_crc)); + + /* ... then for additional cycle data for v2 logs ... */ + if (xfs_sb_version_haslogv2(&log->l_mp->m_sb)) { + union xlog_in_core2 *xhdr = (union xlog_in_core2 *)rhead; + int i; + + for (i = 1; i < log->l_iclog_heads; i++) { + crc = crc32c(crc, &xhdr[i].hic_xheader, + sizeof(struct xlog_rec_ext_header)); + } + } + + /* ... and finally for the payload */ + crc = crc32c(crc, dp, size); + + return xfs_end_cksum(crc); +} /* * The bdstrat callback function for log bufs. This gives us a central @@ -1286,27 +1646,33 @@ xlog_grant_push_ail(xfs_mount_t *mp, * we transition the iclogs to IOERROR state *after* flushing all existing * iclogs to disk. This is because we don't want anymore new transactions to be * started or completed afterwards. + * + * We lock the iclogbufs here so that we can serialise against IO completion + * during unmount. We might be processing a shutdown triggered during unmount, + * and that can occur asynchronously to the unmount thread, and hence we need to + * ensure that completes before tearing down the iclogbufs. Hence we need to + * hold the buffer lock across the log IO to acheive that. */ STATIC int xlog_bdstrat( struct xfs_buf *bp) { - struct xlog_in_core *iclog; + struct xlog_in_core *iclog = bp->b_fspriv; - iclog = XFS_BUF_FSPRIVATE(bp, xlog_in_core_t *); + xfs_buf_lock(bp); if (iclog->ic_state & XLOG_STATE_IOERROR) { - XFS_BUF_ERROR(bp, EIO); - XFS_BUF_STALE(bp); + xfs_buf_ioerror(bp, EIO); + xfs_buf_stale(bp); xfs_buf_ioend(bp, 0); /* * It would seem logical to return EIO here, but we rely on * the log state machine to propagate I/O errors instead of - * doing it here. + * doing it here. Similarly, IO completion will unlock the + * buffer, so we don't do it here. */ return 0; } - bp->b_flags |= _XBF_RUN_QUEUES; xfs_buf_iorequest(bp); return 0; } @@ -1337,10 +1703,10 @@ xlog_bdstrat( */ STATIC int -xlog_sync(xlog_t *log, - xlog_in_core_t *iclog) +xlog_sync( + struct xlog *log, + struct xlog_in_core *iclog) { - xfs_caddr_t dptr; /* pointer to byte sized element */ xfs_buf_t *bp; int i; uint count; /* byte count of bwrite */ @@ -1349,6 +1715,7 @@ xlog_sync(xlog_t *log, int split = 0; /* split write into two regions */ int error; int v2 = xfs_sb_version_haslogv2(&log->l_mp->m_sb); + int size; XFS_STATS_INC(xs_log_writes); ASSERT(atomic_read(&iclog->ic_refcnt) == 0); @@ -1372,51 +1739,83 @@ xlog_sync(xlog_t *log, roundoff < BBTOB(1))); /* move grant heads by roundoff in sync */ - spin_lock(&log->l_grant_lock); - xlog_grant_add_space(log, roundoff); - spin_unlock(&log->l_grant_lock); + xlog_grant_add_space(log, &log->l_reserve_head.grant, roundoff); + xlog_grant_add_space(log, &log->l_write_head.grant, roundoff); /* put cycle number in every block */ xlog_pack_data(log, iclog, roundoff); /* real byte length */ - if (v2) { - iclog->ic_header.h_len = - cpu_to_be32(iclog->ic_offset + roundoff); - } else { - iclog->ic_header.h_len = - cpu_to_be32(iclog->ic_offset); - } + size = iclog->ic_offset; + if (v2) + size += roundoff; + iclog->ic_header.h_len = cpu_to_be32(size); bp = iclog->ic_bp; - ASSERT(XFS_BUF_FSPRIVATE2(bp, unsigned long) == (unsigned long)1); - XFS_BUF_SET_FSPRIVATE2(bp, (unsigned long)2); XFS_BUF_SET_ADDR(bp, BLOCK_LSN(be64_to_cpu(iclog->ic_header.h_lsn))); XFS_STATS_ADD(xs_log_blocks, BTOBB(count)); /* Do we need to split this write into 2 parts? */ if (XFS_BUF_ADDR(bp) + BTOBB(count) > log->l_logBBsize) { + char *dptr; + split = count - (BBTOB(log->l_logBBsize - XFS_BUF_ADDR(bp))); count = BBTOB(log->l_logBBsize - XFS_BUF_ADDR(bp)); - iclog->ic_bwritecnt = 2; /* split into 2 writes */ + iclog->ic_bwritecnt = 2; + + /* + * Bump the cycle numbers at the start of each block in the + * part of the iclog that ends up in the buffer that gets + * written to the start of the log. + * + * Watch out for the header magic number case, though. + */ + dptr = (char *)&iclog->ic_header + count; + for (i = 0; i < split; i += BBSIZE) { + __uint32_t cycle = be32_to_cpu(*(__be32 *)dptr); + if (++cycle == XLOG_HEADER_MAGIC_NUM) + cycle++; + *(__be32 *)dptr = cpu_to_be32(cycle); + + dptr += BBSIZE; + } } else { iclog->ic_bwritecnt = 1; } - XFS_BUF_SET_COUNT(bp, count); - XFS_BUF_SET_FSPRIVATE(bp, iclog); /* save for later */ + + /* calculcate the checksum */ + iclog->ic_header.h_crc = xlog_cksum(log, &iclog->ic_header, + iclog->ic_datap, size); + + bp->b_io_length = BTOBB(count); + bp->b_fspriv = iclog; XFS_BUF_ZEROFLAGS(bp); - XFS_BUF_BUSY(bp); XFS_BUF_ASYNC(bp); - bp->b_flags |= XBF_LOG_BUFFER; + bp->b_flags |= XBF_SYNCIO; - if (log->l_mp->m_flags & XFS_MOUNT_BARRIER) - XFS_BUF_ORDERED(bp); + if (log->l_mp->m_flags & XFS_MOUNT_BARRIER) { + bp->b_flags |= XBF_FUA; + + /* + * Flush the data device before flushing the log to make + * sure all meta data written back from the AIL actually made + * it to disk before stamping the new log tail LSN into the + * log buffer. For an external log we need to issue the + * flush explicitly, and unfortunately synchronously here; + * for an internal log we can simply use the block layer + * state machine for preflushes. + */ + if (log->l_mp->m_logdev_targp != log->l_mp->m_ddev_targp) + xfs_blkdev_issue_flush(log->l_mp->m_ddev_targp); + else + bp->b_flags |= XBF_FLUSH; + } ASSERT(XFS_BUF_ADDR(bp) <= log->l_logBBsize-1); ASSERT(XFS_BUF_ADDR(bp) + BTOBB(count) <= log->l_logBBsize); - xlog_verify_iclog(log, iclog, count, B_TRUE); + xlog_verify_iclog(log, iclog, count, true); /* account for log which doesn't start at block #0 */ XFS_BUF_SET_ADDR(bp, XFS_BUF_ADDR(bp) + log->l_logBBstart); @@ -1426,39 +1825,22 @@ xlog_sync(xlog_t *log, */ XFS_BUF_WRITE(bp); - if ((error = xlog_bdstrat(bp))) { - xfs_ioerror_alert("xlog_sync", log->l_mp, bp, - XFS_BUF_ADDR(bp)); + error = xlog_bdstrat(bp); + if (error) { + xfs_buf_ioerror_alert(bp, "xlog_sync"); return error; } if (split) { bp = iclog->ic_log->l_xbuf; - ASSERT(XFS_BUF_FSPRIVATE2(bp, unsigned long) == - (unsigned long)1); - XFS_BUF_SET_FSPRIVATE2(bp, (unsigned long)2); XFS_BUF_SET_ADDR(bp, 0); /* logical 0 */ - XFS_BUF_SET_PTR(bp, (xfs_caddr_t)((__psint_t)&(iclog->ic_header)+ - (__psint_t)count), split); - XFS_BUF_SET_FSPRIVATE(bp, iclog); + xfs_buf_associate_memory(bp, + (char *)&iclog->ic_header + count, split); + bp->b_fspriv = iclog; XFS_BUF_ZEROFLAGS(bp); - XFS_BUF_BUSY(bp); XFS_BUF_ASYNC(bp); - bp->b_flags |= XBF_LOG_BUFFER; + bp->b_flags |= XBF_SYNCIO; if (log->l_mp->m_flags & XFS_MOUNT_BARRIER) - XFS_BUF_ORDERED(bp); - dptr = XFS_BUF_PTR(bp); - /* - * Bump the cycle numbers at the start of each block - * since this part of the buffer is at the start of - * a new cycle. Watch out for the header magic number - * case, though. - */ - for (i = 0; i < split; i += BBSIZE) { - be32_add_cpu((__be32 *)dptr, 1); - if (be32_to_cpu(*(__be32 *)dptr) == XLOG_HEADER_MAGIC_NUM) - be32_add_cpu((__be32 *)dptr, 1); - dptr += BBSIZE; - } + bp->b_flags |= XBF_FUA; ASSERT(XFS_BUF_ADDR(bp) <= log->l_logBBsize-1); ASSERT(XFS_BUF_ADDR(bp) + BTOBB(count) <= log->l_logBBsize); @@ -1466,40 +1848,57 @@ xlog_sync(xlog_t *log, /* account for internal log which doesn't start at block #0 */ XFS_BUF_SET_ADDR(bp, XFS_BUF_ADDR(bp) + log->l_logBBstart); XFS_BUF_WRITE(bp); - if ((error = xlog_bdstrat(bp))) { - xfs_ioerror_alert("xlog_sync (split)", log->l_mp, - bp, XFS_BUF_ADDR(bp)); + error = xlog_bdstrat(bp); + if (error) { + xfs_buf_ioerror_alert(bp, "xlog_sync (split)"); return error; } } return 0; } /* xlog_sync */ - /* * Deallocate a log structure */ STATIC void -xlog_dealloc_log(xlog_t *log) +xlog_dealloc_log( + struct xlog *log) { xlog_in_core_t *iclog, *next_iclog; int i; xlog_cil_destroy(log); + /* + * Cycle all the iclogbuf locks to make sure all log IO completion + * is done before we tear down these buffers. + */ + iclog = log->l_iclog; + for (i = 0; i < log->l_iclog_bufs; i++) { + xfs_buf_lock(iclog->ic_bp); + xfs_buf_unlock(iclog->ic_bp); + iclog = iclog->ic_next; + } + + /* + * Always need to ensure that the extra buffer does not point to memory + * owned by another log buffer before we free it. Also, cycle the lock + * first to ensure we've completed IO on it. + */ + xfs_buf_lock(log->l_xbuf); + xfs_buf_unlock(log->l_xbuf); + xfs_buf_set_empty(log->l_xbuf, BTOBB(log->l_iclog_size)); + xfs_buf_free(log->l_xbuf); + iclog = log->l_iclog; - for (i=0; i<log->l_iclog_bufs; i++) { - sv_destroy(&iclog->ic_force_wait); - sv_destroy(&iclog->ic_write_wait); + for (i = 0; i < log->l_iclog_bufs; i++) { xfs_buf_free(iclog->ic_bp); next_iclog = iclog->ic_next; kmem_free(iclog); iclog = next_iclog; } spinlock_destroy(&log->l_icloglock); - spinlock_destroy(&log->l_grant_lock); - xfs_buf_free(log->l_xbuf); log->l_mp->m_log = NULL; kmem_free(log); } /* xlog_dealloc_log */ @@ -1509,10 +1908,11 @@ xlog_dealloc_log(xlog_t *log) */ /* ARGSUSED */ static inline void -xlog_state_finish_copy(xlog_t *log, - xlog_in_core_t *iclog, - int record_cnt, - int copy_bytes) +xlog_state_finish_copy( + struct xlog *log, + struct xlog_in_core *iclog, + int record_cnt, + int copy_bytes) { spin_lock(&log->l_icloglock); @@ -1602,40 +2002,38 @@ xlog_print_tic_res( "SWAPEXT" }; - xfs_fs_cmn_err(CE_WARN, mp, - "xfs_log_write: reservation summary:\n" - " trans type = %s (%u)\n" - " unit res = %d bytes\n" - " current res = %d bytes\n" - " total reg = %u bytes (o/flow = %u bytes)\n" - " ophdrs = %u (ophdr space = %u bytes)\n" - " ophdr + reg = %u bytes\n" - " num regions = %u\n", - ((ticket->t_trans_type <= 0 || - ticket->t_trans_type > XFS_TRANS_TYPE_MAX) ? - "bad-trans-type" : trans_type_str[ticket->t_trans_type-1]), - ticket->t_trans_type, - ticket->t_unit_res, - ticket->t_curr_res, - ticket->t_res_arr_sum, ticket->t_res_o_flow, - ticket->t_res_num_ophdrs, ophdr_spc, - ticket->t_res_arr_sum + - ticket->t_res_o_flow + ophdr_spc, - ticket->t_res_num); + xfs_warn(mp, + "xlog_write: reservation summary:\n" + " trans type = %s (%u)\n" + " unit res = %d bytes\n" + " current res = %d bytes\n" + " total reg = %u bytes (o/flow = %u bytes)\n" + " ophdrs = %u (ophdr space = %u bytes)\n" + " ophdr + reg = %u bytes\n" + " num regions = %u\n", + ((ticket->t_trans_type <= 0 || + ticket->t_trans_type > XFS_TRANS_TYPE_MAX) ? + "bad-trans-type" : trans_type_str[ticket->t_trans_type-1]), + ticket->t_trans_type, + ticket->t_unit_res, + ticket->t_curr_res, + ticket->t_res_arr_sum, ticket->t_res_o_flow, + ticket->t_res_num_ophdrs, ophdr_spc, + ticket->t_res_arr_sum + + ticket->t_res_o_flow + ophdr_spc, + ticket->t_res_num); for (i = 0; i < ticket->t_res_num; i++) { - uint r_type = ticket->t_res_arr[i].r_type; - cmn_err(CE_WARN, - "region[%u]: %s - %u bytes\n", - i, + uint r_type = ticket->t_res_arr[i].r_type; + xfs_warn(mp, "region[%u]: %s - %u bytes", i, ((r_type <= 0 || r_type > XLOG_REG_TYPE_MAX) ? "bad-rtype" : res_type_str[r_type-1]), ticket->t_res_arr[i].r_len); } - xfs_cmn_err(XFS_PTAG_LOGRES, CE_ALERT, mp, - "xfs_log_write: reservation ran out. Need to up reservation"); - xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE); + xfs_alert_tag(mp, XFS_PTAG_LOGRES, + "xlog_write: reservation ran out. Need to up reservation"); + xfs_force_shutdown(mp, SHUTDOWN_LOG_IO_ERROR); } /* @@ -1657,6 +2055,10 @@ xlog_write_calc_vec_length( headers++; for (lv = log_vector; lv; lv = lv->lv_next) { + /* we don't write ordered log vectors */ + if (lv->lv_buf_len == XFS_LOG_VEC_ORDERED) + continue; + headers += lv->lv_niovecs; for (i = 0; i < lv->lv_niovecs; i++) { @@ -1698,7 +2100,7 @@ xlog_write_start_rec( static xlog_op_header_t * xlog_write_setup_ophdr( - struct log *log, + struct xlog *log, struct xlog_op_header *ophdr, struct xlog_ticket *ticket, uint flags) @@ -1721,7 +2123,7 @@ xlog_write_setup_ophdr( case XFS_LOG: break; default: - xfs_fs_cmn_err(CE_WARN, log->l_mp, + xfs_warn(log->l_mp, "Bad XFS transaction clientid 0x%x in ticket 0x%p", ophdr->oh_clientid, ticket); return NULL; @@ -1734,7 +2136,7 @@ xlog_write_setup_ophdr( * Set up the parameters of the region copy into the log. This has * to handle region write split across multiple log buffers - this * state is kept external to this function so that this code can - * can be written in an obvious, self documenting manner. + * be written in an obvious, self documenting manner. */ static int xlog_write_setup_copy( @@ -1781,7 +2183,7 @@ xlog_write_setup_copy( static int xlog_write_copy_finish( - struct log *log, + struct xlog *log, struct xlog_in_core *iclog, uint flags, int *record_cnt, @@ -1866,7 +2268,7 @@ xlog_write_copy_finish( */ int xlog_write( - struct log *log, + struct xlog *log, struct xfs_log_vec *log_vector, struct xlog_ticket *ticket, xfs_lsn_t *start_lsn, @@ -1888,23 +2290,21 @@ xlog_write( *start_lsn = 0; len = xlog_write_calc_vec_length(ticket, log_vector); - if (log->l_cilp) { - /* - * Region headers and bytes are already accounted for. - * We only need to take into account start records and - * split regions in this function. - */ - if (ticket->t_flags & XLOG_TIC_INITED) - ticket->t_curr_res -= sizeof(xlog_op_header_t); - /* - * Commit record headers need to be accounted for. These - * come in as separate writes so are easy to detect. - */ - if (flags & (XLOG_COMMIT_TRANS | XLOG_UNMOUNT_TRANS)) - ticket->t_curr_res -= sizeof(xlog_op_header_t); - } else - ticket->t_curr_res -= len; + /* + * Region headers and bytes are already accounted for. + * We only need to take into account start records and + * split regions in this function. + */ + if (ticket->t_flags & XLOG_TIC_INITED) + ticket->t_curr_res -= sizeof(xlog_op_header_t); + + /* + * Commit record headers need to be accounted for. These + * come in as separate writes so are easy to detect. + */ + if (flags & (XLOG_COMMIT_TRANS | XLOG_UNMOUNT_TRANS)) + ticket->t_curr_res -= sizeof(xlog_op_header_t); if (ticket->t_curr_res < 0) xlog_print_tic_res(log->l_mp, ticket); @@ -1912,7 +2312,7 @@ xlog_write( index = 0; lv = log_vector; vecp = lv->lv_iovecp; - while (lv && index < lv->lv_niovecs) { + while (lv && (!lv->lv_niovecs || index < lv->lv_niovecs)) { void *ptr; int log_offset; @@ -1932,13 +2332,22 @@ xlog_write( * This loop writes out as many regions as can fit in the amount * of space which was allocated by xlog_state_get_iclog_space(). */ - while (lv && index < lv->lv_niovecs) { - struct xfs_log_iovec *reg = &vecp[index]; + while (lv && (!lv->lv_niovecs || index < lv->lv_niovecs)) { + struct xfs_log_iovec *reg; struct xlog_op_header *ophdr; int start_rec_copy; int copy_len; int copy_off; + bool ordered = false; + /* ordered log vectors have no regions to write */ + if (lv->lv_buf_len == XFS_LOG_VEC_ORDERED) { + ASSERT(lv->lv_niovecs == 0); + ordered = true; + goto next_lv; + } + + reg = &vecp[index]; ASSERT(reg->i_len % sizeof(__int32_t) == 0); ASSERT((unsigned long)ptr % sizeof(__int32_t) == 0); @@ -1998,12 +2407,13 @@ xlog_write( break; if (++index == lv->lv_niovecs) { +next_lv: lv = lv->lv_next; index = 0; if (lv) vecp = lv->lv_iovecp; } - if (record_cnt == 0) { + if (record_cnt == 0 && ordered == false) { if (!lv) return 0; break; @@ -2039,7 +2449,8 @@ xlog_write( * State Change: DIRTY -> ACTIVE */ STATIC void -xlog_state_clean_log(xlog_t *log) +xlog_state_clean_log( + struct xlog *log) { xlog_in_core_t *iclog; int changed = 0; @@ -2119,7 +2530,7 @@ xlog_state_clean_log(xlog_t *log) STATIC xfs_lsn_t xlog_get_lowest_lsn( - xlog_t *log) + struct xlog *log) { xlog_in_core_t *lsn_log; xfs_lsn_t lowest_lsn, lsn; @@ -2142,9 +2553,9 @@ xlog_get_lowest_lsn( STATIC void xlog_state_do_callback( - xlog_t *log, - int aborted, - xlog_in_core_t *ciclog) + struct xlog *log, + int aborted, + struct xlog_in_core *ciclog) { xlog_in_core_t *iclog; xlog_in_core_t *first_iclog; /* used to know when we've @@ -2232,7 +2643,7 @@ xlog_state_do_callback( lowest_lsn = xlog_get_lowest_lsn(log); if (lowest_lsn && XFS_LSN_CMP(lowest_lsn, - be64_to_cpu(iclog->ic_header.h_lsn)) < 0) { + be64_to_cpu(iclog->ic_header.h_lsn)) < 0) { iclog = iclog->ic_next; continue; /* Leave this iclog for * another thread */ @@ -2240,23 +2651,34 @@ xlog_state_do_callback( iclog->ic_state = XLOG_STATE_CALLBACK; - spin_unlock(&log->l_icloglock); - /* l_last_sync_lsn field protected by - * l_grant_lock. Don't worry about iclog's lsn. - * No one else can be here except us. + /* + * Completion of a iclog IO does not imply that + * a transaction has completed, as transactions + * can be large enough to span many iclogs. We + * cannot change the tail of the log half way + * through a transaction as this may be the only + * transaction in the log and moving th etail to + * point to the middle of it will prevent + * recovery from finding the start of the + * transaction. Hence we should only update the + * last_sync_lsn if this iclog contains + * transaction completion callbacks on it. + * + * We have to do this before we drop the + * icloglock to ensure we are the only one that + * can update it. */ - spin_lock(&log->l_grant_lock); - ASSERT(XFS_LSN_CMP(log->l_last_sync_lsn, - be64_to_cpu(iclog->ic_header.h_lsn)) <= 0); - log->l_last_sync_lsn = - be64_to_cpu(iclog->ic_header.h_lsn); - spin_unlock(&log->l_grant_lock); + ASSERT(XFS_LSN_CMP(atomic64_read(&log->l_last_sync_lsn), + be64_to_cpu(iclog->ic_header.h_lsn)) <= 0); + if (iclog->ic_callback) + atomic64_set(&log->l_last_sync_lsn, + be64_to_cpu(iclog->ic_header.h_lsn)); - } else { - spin_unlock(&log->l_icloglock); + } else ioerrors++; - } + + spin_unlock(&log->l_icloglock); /* * Keep processing entries in the callback list until @@ -2297,7 +2719,7 @@ xlog_state_do_callback( xlog_state_clean_log(log); /* wake up threads waiting in xfs_log_force() */ - sv_broadcast(&iclog->ic_force_wait); + wake_up_all(&iclog->ic_force_wait); iclog = iclog->ic_next; } while (first_iclog != iclog); @@ -2305,7 +2727,7 @@ xlog_state_do_callback( if (repeats > 5000) { flushcnt += repeats; repeats = 0; - xfs_fs_cmn_err(CE_WARN, log->l_mp, + xfs_warn(log->l_mp, "%s: possible infinite loop (%d iterations)", __func__, flushcnt); } @@ -2344,7 +2766,7 @@ xlog_state_do_callback( spin_unlock(&log->l_icloglock); if (wake) - sv_broadcast(&log->l_flush_wait); + wake_up_all(&log->l_flush_wait); } @@ -2366,7 +2788,7 @@ xlog_state_done_syncing( xlog_in_core_t *iclog, int aborted) { - xlog_t *log = iclog->ic_log; + struct xlog *log = iclog->ic_log; spin_lock(&log->l_icloglock); @@ -2395,7 +2817,7 @@ xlog_state_done_syncing( * iclog buffer, we wake them all, one will get to do the * I/O, the others get to wait for the result. */ - sv_broadcast(&iclog->ic_write_wait); + wake_up_all(&iclog->ic_write_wait); spin_unlock(&log->l_icloglock); xlog_state_do_callback(log, aborted, iclog); /* also cleans log */ } /* xlog_state_done_syncing */ @@ -2420,12 +2842,13 @@ xlog_state_done_syncing( * is copied. */ STATIC int -xlog_state_get_iclog_space(xlog_t *log, - int len, - xlog_in_core_t **iclogp, - xlog_ticket_t *ticket, - int *continued_write, - int *logoffsetp) +xlog_state_get_iclog_space( + struct xlog *log, + int len, + struct xlog_in_core **iclogp, + struct xlog_ticket *ticket, + int *continued_write, + int *logoffsetp) { int log_offset; xlog_rec_header_t *head; @@ -2444,7 +2867,7 @@ restart: XFS_STATS_INC(xs_log_noiclogs); /* Wait for log writes to have flushed */ - sv_wait(&log->l_flush_wait, 0, &log->l_icloglock, 0); + xlog_wait(&log->l_flush_wait, &log->l_icloglock); goto restart; } @@ -2522,267 +2945,6 @@ restart: return 0; } /* xlog_state_get_iclog_space */ -/* - * Atomically get the log space required for a log ticket. - * - * Once a ticket gets put onto the reserveq, it will only return after - * the needed reservation is satisfied. - */ -STATIC int -xlog_grant_log_space(xlog_t *log, - xlog_ticket_t *tic) -{ - int free_bytes; - int need_bytes; -#ifdef DEBUG - xfs_lsn_t tail_lsn; -#endif - - -#ifdef DEBUG - if (log->l_flags & XLOG_ACTIVE_RECOVERY) - panic("grant Recovery problem"); -#endif - - /* Is there space or do we need to sleep? */ - spin_lock(&log->l_grant_lock); - - trace_xfs_log_grant_enter(log, tic); - - /* something is already sleeping; insert new transaction at end */ - if (log->l_reserve_headq) { - xlog_ins_ticketq(&log->l_reserve_headq, tic); - - trace_xfs_log_grant_sleep1(log, tic); - - /* - * Gotta check this before going to sleep, while we're - * holding the grant lock. - */ - if (XLOG_FORCED_SHUTDOWN(log)) - goto error_return; - - XFS_STATS_INC(xs_sleep_logspace); - sv_wait(&tic->t_wait, PINOD|PLTWAIT, &log->l_grant_lock, s); - /* - * If we got an error, and the filesystem is shutting down, - * we'll catch it down below. So just continue... - */ - trace_xfs_log_grant_wake1(log, tic); - spin_lock(&log->l_grant_lock); - } - if (tic->t_flags & XFS_LOG_PERM_RESERV) - need_bytes = tic->t_unit_res*tic->t_ocnt; - else - need_bytes = tic->t_unit_res; - -redo: - if (XLOG_FORCED_SHUTDOWN(log)) - goto error_return; - - free_bytes = xlog_space_left(log, log->l_grant_reserve_cycle, - log->l_grant_reserve_bytes); - if (free_bytes < need_bytes) { - if ((tic->t_flags & XLOG_TIC_IN_Q) == 0) - xlog_ins_ticketq(&log->l_reserve_headq, tic); - - trace_xfs_log_grant_sleep2(log, tic); - - spin_unlock(&log->l_grant_lock); - xlog_grant_push_ail(log->l_mp, need_bytes); - spin_lock(&log->l_grant_lock); - - XFS_STATS_INC(xs_sleep_logspace); - sv_wait(&tic->t_wait, PINOD|PLTWAIT, &log->l_grant_lock, s); - - spin_lock(&log->l_grant_lock); - if (XLOG_FORCED_SHUTDOWN(log)) - goto error_return; - - trace_xfs_log_grant_wake2(log, tic); - - goto redo; - } else if (tic->t_flags & XLOG_TIC_IN_Q) - xlog_del_ticketq(&log->l_reserve_headq, tic); - - /* we've got enough space */ - xlog_grant_add_space(log, need_bytes); -#ifdef DEBUG - tail_lsn = log->l_tail_lsn; - /* - * Check to make sure the grant write head didn't just over lap the - * tail. If the cycles are the same, we can't be overlapping. - * Otherwise, make sure that the cycles differ by exactly one and - * check the byte count. - */ - if (CYCLE_LSN(tail_lsn) != log->l_grant_write_cycle) { - ASSERT(log->l_grant_write_cycle-1 == CYCLE_LSN(tail_lsn)); - ASSERT(log->l_grant_write_bytes <= BBTOB(BLOCK_LSN(tail_lsn))); - } -#endif - trace_xfs_log_grant_exit(log, tic); - xlog_verify_grant_head(log, 1); - spin_unlock(&log->l_grant_lock); - return 0; - - error_return: - if (tic->t_flags & XLOG_TIC_IN_Q) - xlog_del_ticketq(&log->l_reserve_headq, tic); - - trace_xfs_log_grant_error(log, tic); - - /* - * If we are failing, make sure the ticket doesn't have any - * current reservations. We don't want to add this back when - * the ticket/transaction gets cancelled. - */ - tic->t_curr_res = 0; - tic->t_cnt = 0; /* ungrant will give back unit_res * t_cnt. */ - spin_unlock(&log->l_grant_lock); - return XFS_ERROR(EIO); -} /* xlog_grant_log_space */ - - -/* - * Replenish the byte reservation required by moving the grant write head. - * - * - */ -STATIC int -xlog_regrant_write_log_space(xlog_t *log, - xlog_ticket_t *tic) -{ - int free_bytes, need_bytes; - xlog_ticket_t *ntic; -#ifdef DEBUG - xfs_lsn_t tail_lsn; -#endif - - tic->t_curr_res = tic->t_unit_res; - xlog_tic_reset_res(tic); - - if (tic->t_cnt > 0) - return 0; - -#ifdef DEBUG - if (log->l_flags & XLOG_ACTIVE_RECOVERY) - panic("regrant Recovery problem"); -#endif - - spin_lock(&log->l_grant_lock); - - trace_xfs_log_regrant_write_enter(log, tic); - - if (XLOG_FORCED_SHUTDOWN(log)) - goto error_return; - - /* If there are other waiters on the queue then give them a - * chance at logspace before us. Wake up the first waiters, - * if we do not wake up all the waiters then go to sleep waiting - * for more free space, otherwise try to get some space for - * this transaction. - */ - need_bytes = tic->t_unit_res; - if ((ntic = log->l_write_headq)) { - free_bytes = xlog_space_left(log, log->l_grant_write_cycle, - log->l_grant_write_bytes); - do { - ASSERT(ntic->t_flags & XLOG_TIC_PERM_RESERV); - - if (free_bytes < ntic->t_unit_res) - break; - free_bytes -= ntic->t_unit_res; - sv_signal(&ntic->t_wait); - ntic = ntic->t_next; - } while (ntic != log->l_write_headq); - - if (ntic != log->l_write_headq) { - if ((tic->t_flags & XLOG_TIC_IN_Q) == 0) - xlog_ins_ticketq(&log->l_write_headq, tic); - - trace_xfs_log_regrant_write_sleep1(log, tic); - - spin_unlock(&log->l_grant_lock); - xlog_grant_push_ail(log->l_mp, need_bytes); - spin_lock(&log->l_grant_lock); - - XFS_STATS_INC(xs_sleep_logspace); - sv_wait(&tic->t_wait, PINOD|PLTWAIT, - &log->l_grant_lock, s); - - /* If we're shutting down, this tic is already - * off the queue */ - spin_lock(&log->l_grant_lock); - if (XLOG_FORCED_SHUTDOWN(log)) - goto error_return; - - trace_xfs_log_regrant_write_wake1(log, tic); - } - } - -redo: - if (XLOG_FORCED_SHUTDOWN(log)) - goto error_return; - - free_bytes = xlog_space_left(log, log->l_grant_write_cycle, - log->l_grant_write_bytes); - if (free_bytes < need_bytes) { - if ((tic->t_flags & XLOG_TIC_IN_Q) == 0) - xlog_ins_ticketq(&log->l_write_headq, tic); - spin_unlock(&log->l_grant_lock); - xlog_grant_push_ail(log->l_mp, need_bytes); - spin_lock(&log->l_grant_lock); - - XFS_STATS_INC(xs_sleep_logspace); - trace_xfs_log_regrant_write_sleep2(log, tic); - - sv_wait(&tic->t_wait, PINOD|PLTWAIT, &log->l_grant_lock, s); - - /* If we're shutting down, this tic is already off the queue */ - spin_lock(&log->l_grant_lock); - if (XLOG_FORCED_SHUTDOWN(log)) - goto error_return; - - trace_xfs_log_regrant_write_wake2(log, tic); - goto redo; - } else if (tic->t_flags & XLOG_TIC_IN_Q) - xlog_del_ticketq(&log->l_write_headq, tic); - - /* we've got enough space */ - xlog_grant_add_space_write(log, need_bytes); -#ifdef DEBUG - tail_lsn = log->l_tail_lsn; - if (CYCLE_LSN(tail_lsn) != log->l_grant_write_cycle) { - ASSERT(log->l_grant_write_cycle-1 == CYCLE_LSN(tail_lsn)); - ASSERT(log->l_grant_write_bytes <= BBTOB(BLOCK_LSN(tail_lsn))); - } -#endif - - trace_xfs_log_regrant_write_exit(log, tic); - - xlog_verify_grant_head(log, 1); - spin_unlock(&log->l_grant_lock); - return 0; - - - error_return: - if (tic->t_flags & XLOG_TIC_IN_Q) - xlog_del_ticketq(&log->l_reserve_headq, tic); - - trace_xfs_log_regrant_write_error(log, tic); - - /* - * If we are failing, make sure the ticket doesn't have any - * current reservations. We don't want to add this back when - * the ticket/transaction gets cancelled. - */ - tic->t_curr_res = 0; - tic->t_cnt = 0; /* ungrant will give back unit_res * t_cnt. */ - spin_unlock(&log->l_grant_lock); - return XFS_ERROR(EIO); -} /* xlog_regrant_write_log_space */ - - /* The first cnt-1 times through here we don't need to * move the grant write head because the permanent * reservation has reserved cnt times the unit amount. @@ -2791,35 +2953,33 @@ redo: * move grant reservation head forward. */ STATIC void -xlog_regrant_reserve_log_space(xlog_t *log, - xlog_ticket_t *ticket) +xlog_regrant_reserve_log_space( + struct xlog *log, + struct xlog_ticket *ticket) { trace_xfs_log_regrant_reserve_enter(log, ticket); if (ticket->t_cnt > 0) ticket->t_cnt--; - spin_lock(&log->l_grant_lock); - xlog_grant_sub_space(log, ticket->t_curr_res); + xlog_grant_sub_space(log, &log->l_reserve_head.grant, + ticket->t_curr_res); + xlog_grant_sub_space(log, &log->l_write_head.grant, + ticket->t_curr_res); ticket->t_curr_res = ticket->t_unit_res; xlog_tic_reset_res(ticket); trace_xfs_log_regrant_reserve_sub(log, ticket); - xlog_verify_grant_head(log, 1); - /* just return if we still have some of the pre-reserved space */ - if (ticket->t_cnt > 0) { - spin_unlock(&log->l_grant_lock); + if (ticket->t_cnt > 0) return; - } - xlog_grant_add_space_reserve(log, ticket->t_unit_res); + xlog_grant_add_space(log, &log->l_reserve_head.grant, + ticket->t_unit_res); trace_xfs_log_regrant_reserve_exit(log, ticket); - xlog_verify_grant_head(log, 0); - spin_unlock(&log->l_grant_lock); ticket->t_curr_res = ticket->t_unit_res; xlog_tic_reset_res(ticket); } /* xlog_regrant_reserve_log_space */ @@ -2840,34 +3000,35 @@ xlog_regrant_reserve_log_space(xlog_t *log, * in the current reservation field. */ STATIC void -xlog_ungrant_log_space(xlog_t *log, - xlog_ticket_t *ticket) +xlog_ungrant_log_space( + struct xlog *log, + struct xlog_ticket *ticket) { + int bytes; + if (ticket->t_cnt > 0) ticket->t_cnt--; - spin_lock(&log->l_grant_lock); trace_xfs_log_ungrant_enter(log, ticket); - - xlog_grant_sub_space(log, ticket->t_curr_res); - trace_xfs_log_ungrant_sub(log, ticket); - /* If this is a permanent reservation ticket, we may be able to free + /* + * If this is a permanent reservation ticket, we may be able to free * up more space based on the remaining count. */ + bytes = ticket->t_curr_res; if (ticket->t_cnt > 0) { ASSERT(ticket->t_flags & XLOG_TIC_PERM_RESERV); - xlog_grant_sub_space(log, ticket->t_unit_res*ticket->t_cnt); + bytes += ticket->t_unit_res*ticket->t_cnt; } - trace_xfs_log_ungrant_exit(log, ticket); + xlog_grant_sub_space(log, &log->l_reserve_head.grant, bytes); + xlog_grant_sub_space(log, &log->l_write_head.grant, bytes); - xlog_verify_grant_head(log, 1); - spin_unlock(&log->l_grant_lock); - xfs_log_move_tail(log->l_mp, 1); -} /* xlog_ungrant_log_space */ + trace_xfs_log_ungrant_exit(log, ticket); + xfs_log_space_wake(log->l_mp); +} /* * Flush iclog to disk if this is the last reference to the given iclog and @@ -2880,8 +3041,8 @@ xlog_ungrant_log_space(xlog_t *log, */ STATIC int xlog_state_release_iclog( - xlog_t *log, - xlog_in_core_t *iclog) + struct xlog *log, + struct xlog_in_core *iclog) { int sync = 0; /* do we sync? */ @@ -2901,11 +3062,11 @@ xlog_state_release_iclog( if (iclog->ic_state == XLOG_STATE_WANT_SYNC) { /* update tail before writing to iclog */ - xlog_assign_tail_lsn(log->l_mp); + xfs_lsn_t tail_lsn = xlog_assign_tail_lsn(log->l_mp); sync++; iclog->ic_state = XLOG_STATE_SYNCING; - iclog->ic_header.h_tail_lsn = cpu_to_be64(log->l_tail_lsn); - xlog_verify_tail_lsn(log, iclog, log->l_tail_lsn); + iclog->ic_header.h_tail_lsn = cpu_to_be64(tail_lsn); + xlog_verify_tail_lsn(log, iclog, tail_lsn); /* cycle incremented when incrementing curr_block */ } spin_unlock(&log->l_icloglock); @@ -2931,9 +3092,10 @@ xlog_state_release_iclog( * that every data block. We have run out of space in this log record. */ STATIC void -xlog_state_switch_iclogs(xlog_t *log, - xlog_in_core_t *iclog, - int eventual_size) +xlog_state_switch_iclogs( + struct xlog *log, + struct xlog_in_core *iclog, + int eventual_size) { ASSERT(iclog->ic_state == XLOG_STATE_ACTIVE); if (!eventual_size) @@ -2997,14 +3159,13 @@ _xfs_log_force( uint flags, int *log_flushed) { - struct log *log = mp->m_log; + struct xlog *log = mp->m_log; struct xlog_in_core *iclog; xfs_lsn_t lsn; XFS_STATS_INC(xs_log_force); - if (log->l_cilp) - xlog_cil_force(log); + xlog_cil_force(log); spin_lock(&log->l_icloglock); @@ -3088,7 +3249,7 @@ maybe_sleep: return XFS_ERROR(EIO); } XFS_STATS_INC(xs_log_force_sleep); - sv_wait(&iclog->ic_force_wait, PINOD, &log->l_icloglock, s); + xlog_wait(&iclog->ic_force_wait, &log->l_icloglock); /* * No need to grab the log lock here since we're * only deciding whether or not to return EIO @@ -3118,11 +3279,10 @@ xfs_log_force( { int error; + trace_xfs_log_force(mp, 0); error = _xfs_log_force(mp, flags, NULL); - if (error) { - xfs_fs_cmn_err(CE_WARN, mp, "xfs_log_force: " - "error %d returned.", error); - } + if (error) + xfs_warn(mp, "%s: error %d returned.", __func__, error); } /* @@ -3147,7 +3307,7 @@ _xfs_log_force_lsn( uint flags, int *log_flushed) { - struct log *log = mp->m_log; + struct xlog *log = mp->m_log; struct xlog_in_core *iclog; int already_slept = 0; @@ -3155,11 +3315,9 @@ _xfs_log_force_lsn( XFS_STATS_INC(xs_log_force); - if (log->l_cilp) { - lsn = xlog_cil_force_lsn(log, lsn); - if (lsn == NULLCOMMITLSN) - return 0; - } + lsn = xlog_cil_force_lsn(log, lsn); + if (lsn == NULLCOMMITLSN) + return 0; try_again: spin_lock(&log->l_icloglock); @@ -3206,8 +3364,8 @@ try_again: XFS_STATS_INC(xs_log_force_sleep); - sv_wait(&iclog->ic_prev->ic_write_wait, - PSWP, &log->l_icloglock, s); + xlog_wait(&iclog->ic_prev->ic_write_wait, + &log->l_icloglock); if (log_flushed) *log_flushed = 1; already_slept = 1; @@ -3235,7 +3393,7 @@ try_again: return XFS_ERROR(EIO); } XFS_STATS_INC(xs_log_force_sleep); - sv_wait(&iclog->ic_force_wait, PSWP, &log->l_icloglock, s); + xlog_wait(&iclog->ic_force_wait, &log->l_icloglock); /* * No need to grab the log lock here since we're * only deciding whether or not to return EIO @@ -3270,11 +3428,10 @@ xfs_log_force_lsn( { int error; + trace_xfs_log_force(mp, lsn); error = _xfs_log_force_lsn(mp, lsn, flags, NULL); - if (error) { - xfs_fs_cmn_err(CE_WARN, mp, "xfs_log_force: " - "error %d returned.", error); - } + if (error) + xfs_warn(mp, "%s: error %d returned.", __func__, error); } /* @@ -3282,7 +3439,9 @@ xfs_log_force_lsn( * disk. */ STATIC void -xlog_state_want_sync(xlog_t *log, xlog_in_core_t *iclog) +xlog_state_want_sync( + struct xlog *log, + struct xlog_in_core *iclog) { assert_spin_locked(&log->l_icloglock); @@ -3310,10 +3469,8 @@ xfs_log_ticket_put( xlog_ticket_t *ticket) { ASSERT(atomic_read(&ticket->t_ref) > 0); - if (atomic_dec_and_test(&ticket->t_ref)) { - sv_destroy(&ticket->t_wait); + if (atomic_dec_and_test(&ticket->t_ref)) kmem_zone_free(xfs_log_ticket_zone, ticket); - } } xlog_ticket_t * @@ -3325,32 +3482,18 @@ xfs_log_ticket_get( return ticket; } -xlog_tid_t -xfs_log_get_trans_ident( - struct xfs_trans *tp) -{ - return tp->t_ticket->t_tid; -} - /* - * Allocate and initialise a new log ticket. + * Figure out the total log space unit (in bytes) that would be + * required for a log ticket. */ -xlog_ticket_t * -xlog_ticket_alloc( - struct log *log, - int unit_bytes, - int cnt, - char client, - uint xflags, - int alloc_flags) +int +xfs_log_calc_unit_res( + struct xfs_mount *mp, + int unit_bytes) { - struct xlog_ticket *tic; - uint num_headers; - int iclog_space; - - tic = kmem_zone_zalloc(xfs_log_ticket_zone, alloc_flags); - if (!tic) - return NULL; + struct xlog *log = mp->m_log; + int iclog_space; + uint num_headers; /* * Permanent reservations have up to 'cnt'-1 active log operations @@ -3425,27 +3568,51 @@ xlog_ticket_alloc( unit_bytes += log->l_iclog_hsize; /* for roundoff padding for transaction data and one for commit record */ - if (xfs_sb_version_haslogv2(&log->l_mp->m_sb) && - log->l_mp->m_sb.sb_logsunit > 1) { + if (xfs_sb_version_haslogv2(&mp->m_sb) && mp->m_sb.sb_logsunit > 1) { /* log su roundoff */ - unit_bytes += 2*log->l_mp->m_sb.sb_logsunit; + unit_bytes += 2 * mp->m_sb.sb_logsunit; } else { /* BB roundoff */ - unit_bytes += 2*BBSIZE; + unit_bytes += 2 * BBSIZE; } + return unit_bytes; +} + +/* + * Allocate and initialise a new log ticket. + */ +struct xlog_ticket * +xlog_ticket_alloc( + struct xlog *log, + int unit_bytes, + int cnt, + char client, + bool permanent, + xfs_km_flags_t alloc_flags) +{ + struct xlog_ticket *tic; + int unit_res; + + tic = kmem_zone_zalloc(xfs_log_ticket_zone, alloc_flags); + if (!tic) + return NULL; + + unit_res = xfs_log_calc_unit_res(log->l_mp, unit_bytes); + atomic_set(&tic->t_ref, 1); - tic->t_unit_res = unit_bytes; - tic->t_curr_res = unit_bytes; + tic->t_task = current; + INIT_LIST_HEAD(&tic->t_queue); + tic->t_unit_res = unit_res; + tic->t_curr_res = unit_res; tic->t_cnt = cnt; tic->t_ocnt = cnt; - tic->t_tid = random32(); + tic->t_tid = prandom_u32(); tic->t_clientid = client; tic->t_flags = XLOG_TIC_INITED; tic->t_trans_type = 0; - if (xflags & XFS_LOG_PERM_RESERV) + if (permanent) tic->t_flags |= XLOG_TIC_PERM_RESERV; - sv_init(&tic->t_wait, SV_DEFAULT, "logtick"); xlog_tic_reset_res(tic); @@ -3467,7 +3634,7 @@ xlog_ticket_alloc( */ void xlog_verify_dest_ptr( - struct log *log, + struct xlog *log, char *ptr) { int i; @@ -3480,28 +3647,52 @@ xlog_verify_dest_ptr( } if (!good_ptr) - xlog_panic("xlog_verify_dest_ptr: invalid ptr"); + xfs_emerg(log->l_mp, "%s: invalid ptr", __func__); } +/* + * Check to make sure the grant write head didn't just over lap the tail. If + * the cycles are the same, we can't be overlapping. Otherwise, make sure that + * the cycles differ by exactly one and check the byte count. + * + * This check is run unlocked, so can give false positives. Rather than assert + * on failures, use a warn-once flag and a panic tag to allow the admin to + * determine if they want to panic the machine when such an error occurs. For + * debug kernels this will have the same effect as using an assert but, unlinke + * an assert, it can be turned off at runtime. + */ STATIC void -xlog_verify_grant_head(xlog_t *log, int equals) -{ - if (log->l_grant_reserve_cycle == log->l_grant_write_cycle) { - if (equals) - ASSERT(log->l_grant_reserve_bytes >= log->l_grant_write_bytes); - else - ASSERT(log->l_grant_reserve_bytes > log->l_grant_write_bytes); - } else { - ASSERT(log->l_grant_reserve_cycle-1 == log->l_grant_write_cycle); - ASSERT(log->l_grant_write_bytes >= log->l_grant_reserve_bytes); - } -} /* xlog_verify_grant_head */ +xlog_verify_grant_tail( + struct xlog *log) +{ + int tail_cycle, tail_blocks; + int cycle, space; + + xlog_crack_grant_head(&log->l_write_head.grant, &cycle, &space); + xlog_crack_atomic_lsn(&log->l_tail_lsn, &tail_cycle, &tail_blocks); + if (tail_cycle != cycle) { + if (cycle - 1 != tail_cycle && + !(log->l_flags & XLOG_TAIL_WARN)) { + xfs_alert_tag(log->l_mp, XFS_PTAG_LOGRES, + "%s: cycle - 1 != tail_cycle", __func__); + log->l_flags |= XLOG_TAIL_WARN; + } + + if (space > BBTOB(tail_blocks) && + !(log->l_flags & XLOG_TAIL_WARN)) { + xfs_alert_tag(log->l_mp, XFS_PTAG_LOGRES, + "%s: space > BBTOB(tail_blocks)", __func__); + log->l_flags |= XLOG_TAIL_WARN; + } + } +} /* check if it will fit */ STATIC void -xlog_verify_tail_lsn(xlog_t *log, - xlog_in_core_t *iclog, - xfs_lsn_t tail_lsn) +xlog_verify_tail_lsn( + struct xlog *log, + struct xlog_in_core *iclog, + xfs_lsn_t tail_lsn) { int blocks; @@ -3509,16 +3700,16 @@ xlog_verify_tail_lsn(xlog_t *log, blocks = log->l_logBBsize - (log->l_prev_block - BLOCK_LSN(tail_lsn)); if (blocks < BTOBB(iclog->ic_offset)+BTOBB(log->l_iclog_hsize)) - xlog_panic("xlog_verify_tail_lsn: ran out of log space"); + xfs_emerg(log->l_mp, "%s: ran out of log space", __func__); } else { ASSERT(CYCLE_LSN(tail_lsn)+1 == log->l_prev_cycle); if (BLOCK_LSN(tail_lsn) == log->l_prev_block) - xlog_panic("xlog_verify_tail_lsn: tail wrapped"); + xfs_emerg(log->l_mp, "%s: tail wrapped", __func__); blocks = BLOCK_LSN(tail_lsn) - log->l_prev_block; if (blocks < BTOBB(iclog->ic_offset) + 1) - xlog_panic("xlog_verify_tail_lsn: ran out of log space"); + xfs_emerg(log->l_mp, "%s: ran out of log space", __func__); } } /* xlog_verify_tail_lsn */ @@ -3538,10 +3729,11 @@ xlog_verify_tail_lsn(xlog_t *log, * the cycle numbers agree with the current cycle number. */ STATIC void -xlog_verify_iclog(xlog_t *log, - xlog_in_core_t *iclog, - int count, - boolean_t syncing) +xlog_verify_iclog( + struct xlog *log, + struct xlog_in_core *iclog, + int count, + bool syncing) { xlog_op_header_t *ophead; xlog_in_core_t *icptr; @@ -3556,24 +3748,23 @@ xlog_verify_iclog(xlog_t *log, /* check validity of iclog pointers */ spin_lock(&log->l_icloglock); icptr = log->l_iclog; - for (i=0; i < log->l_iclog_bufs; i++) { - if (icptr == NULL) - xlog_panic("xlog_verify_iclog: invalid ptr"); - icptr = icptr->ic_next; - } + for (i = 0; i < log->l_iclog_bufs; i++, icptr = icptr->ic_next) + ASSERT(icptr); + if (icptr != log->l_iclog) - xlog_panic("xlog_verify_iclog: corrupt iclog ring"); + xfs_emerg(log->l_mp, "%s: corrupt iclog ring", __func__); spin_unlock(&log->l_icloglock); /* check log magic numbers */ - if (be32_to_cpu(iclog->ic_header.h_magicno) != XLOG_HEADER_MAGIC_NUM) - xlog_panic("xlog_verify_iclog: invalid magic num"); + if (iclog->ic_header.h_magicno != cpu_to_be32(XLOG_HEADER_MAGIC_NUM)) + xfs_emerg(log->l_mp, "%s: invalid magic num", __func__); ptr = (xfs_caddr_t) &iclog->ic_header; for (ptr += BBSIZE; ptr < ((xfs_caddr_t)&iclog->ic_header) + count; ptr += BBSIZE) { - if (be32_to_cpu(*(__be32 *)ptr) == XLOG_HEADER_MAGIC_NUM) - xlog_panic("xlog_verify_iclog: unexpected magic num"); + if (*(__be32 *)ptr == cpu_to_be32(XLOG_HEADER_MAGIC_NUM)) + xfs_emerg(log->l_mp, "%s: unexpected magic num", + __func__); } /* check fields */ @@ -3588,7 +3779,7 @@ xlog_verify_iclog(xlog_t *log, /* clientid is only 1 byte */ field_offset = (__psint_t) ((xfs_caddr_t)&(ophead->oh_clientid) - base_ptr); - if (syncing == B_FALSE || (field_offset & 0x1ff)) { + if (!syncing || (field_offset & 0x1ff)) { clientid = ophead->oh_clientid; } else { idx = BTOBBT((xfs_caddr_t)&(ophead->oh_clientid) - iclog->ic_datap); @@ -3603,14 +3794,15 @@ xlog_verify_iclog(xlog_t *log, } } if (clientid != XFS_TRANSACTION && clientid != XFS_LOG) - cmn_err(CE_WARN, "xlog_verify_iclog: " - "invalid clientid %d op 0x%p offset 0x%lx", - clientid, ophead, (unsigned long)field_offset); + xfs_warn(log->l_mp, + "%s: invalid clientid %d op 0x%p offset 0x%lx", + __func__, clientid, ophead, + (unsigned long)field_offset); /* check length */ field_offset = (__psint_t) ((xfs_caddr_t)&(ophead->oh_len) - base_ptr); - if (syncing == B_FALSE || (field_offset & 0x1ff)) { + if (!syncing || (field_offset & 0x1ff)) { op_len = be32_to_cpu(ophead->oh_len); } else { idx = BTOBBT((__psint_t)&ophead->oh_len - @@ -3633,7 +3825,7 @@ xlog_verify_iclog(xlog_t *log, */ STATIC int xlog_state_ioerror( - xlog_t *log) + struct xlog *log) { xlog_in_core_t *iclog, *ic; @@ -3678,8 +3870,7 @@ xfs_log_force_umount( struct xfs_mount *mp, int logerror) { - xlog_ticket_t *tic; - xlog_t *log; + struct xlog *log; int retval; log = mp->m_log; @@ -3712,16 +3903,14 @@ xfs_log_force_umount( * completed transactions are flushed to disk with the xfs_log_force() * call below. */ - if (!logerror && (mp->m_flags & XFS_MOUNT_DELAYLOG)) + if (!logerror) xlog_cil_force(log); /* - * We must hold both the GRANT lock and the LOG lock, - * before we mark the filesystem SHUTDOWN and wake - * everybody up to tell the bad news. + * mark the filesystem and the as in a shutdown state and wake + * everybody up to tell them the bad news. */ spin_lock(&log->l_icloglock); - spin_lock(&log->l_grant_lock); mp->m_flags |= XFS_MOUNT_FS_SHUTDOWN; if (mp->m_sb_bp) XFS_BUF_DONE(mp->m_sb_bp); @@ -3742,27 +3931,14 @@ xfs_log_force_umount( spin_unlock(&log->l_icloglock); /* - * We don't want anybody waiting for log reservations - * after this. That means we have to wake up everybody - * queued up on reserve_headq as well as write_headq. - * In addition, we make sure in xlog_{re}grant_log_space - * that we don't enqueue anything once the SHUTDOWN flag - * is set, and this action is protected by the GRANTLOCK. + * We don't want anybody waiting for log reservations after this. That + * means we have to wake up everybody queued up on reserveq as well as + * writeq. In addition, we make sure in xlog_{re}grant_log_space that + * we don't enqueue anything once the SHUTDOWN flag is set, and this + * action is protected by the grant locks. */ - if ((tic = log->l_reserve_headq)) { - do { - sv_signal(&tic->t_wait); - tic = tic->t_next; - } while (tic != log->l_reserve_headq); - } - - if ((tic = log->l_write_headq)) { - do { - sv_signal(&tic->t_wait); - tic = tic->t_next; - } while (tic != log->l_write_headq); - } - spin_unlock(&log->l_grant_lock); + xlog_grant_head_wake_all(&log->l_reserve_head); + xlog_grant_head_wake_all(&log->l_write_head); if (!(log->l_iclog->ic_state & XLOG_STATE_IOERROR)) { ASSERT(!logerror); @@ -3776,11 +3952,14 @@ xfs_log_force_umount( retval = xlog_state_ioerror(log); spin_unlock(&log->l_icloglock); } + /* - * Wake up everybody waiting on xfs_log_force. - * Callback all log item committed functions as if the - * log writes were completed. + * Wake up everybody waiting on xfs_log_force. Wake the CIL push first + * as if the log writes were completed. The abort handling in the log + * item committed callback functions will do this again under lock to + * avoid races. */ + wake_up_all(&log->l_cilp->xc_commit_wait); xlog_state_do_callback(log, XFS_LI_ABORTED, NULL); #ifdef XFSERRORDEBUG @@ -3801,7 +3980,8 @@ xfs_log_force_umount( } STATIC int -xlog_iclogs_empty(xlog_t *log) +xlog_iclogs_empty( + struct xlog *log) { xlog_in_core_t *iclog; @@ -3816,3 +3996,4 @@ xlog_iclogs_empty(xlog_t *log) } while (iclog != log->l_iclog); return 1; } + |