diff options
Diffstat (limited to 'fs/xfs/xfs_log.c')
| -rw-r--r-- | fs/xfs/xfs_log.c | 133 |
1 files changed, 102 insertions, 31 deletions
diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c index 3731f13f63e9..dc95a49d62e7 100644 --- a/fs/xfs/xfs_log.c +++ b/fs/xfs/xfs_log.c @@ -434,7 +434,7 @@ xfs_log_reserve( int unit_bytes, int cnt, struct xlog_ticket **ticp, - __uint8_t client, + uint8_t client, bool permanent) { struct xlog *log = mp->m_log; @@ -743,15 +743,45 @@ xfs_log_mount_finish( struct xfs_mount *mp) { int error = 0; + bool readonly = (mp->m_flags & XFS_MOUNT_RDONLY); if (mp->m_flags & XFS_MOUNT_NORECOVERY) { ASSERT(mp->m_flags & XFS_MOUNT_RDONLY); return 0; + } else if (readonly) { + /* Allow unlinked processing to proceed */ + mp->m_flags &= ~XFS_MOUNT_RDONLY; } + /* + * During the second phase of log recovery, we need iget and + * iput to behave like they do for an active filesystem. + * xfs_fs_drop_inode needs to be able to prevent the deletion + * of inodes before we're done replaying log items on those + * inodes. Turn it off immediately after recovery finishes + * so that we don't leak the quota inodes if subsequent mount + * activities fail. + * + * We let all inodes involved in redo item processing end up on + * the LRU instead of being evicted immediately so that if we do + * something to an unlinked inode, the irele won't cause + * premature truncation and freeing of the inode, which results + * in log recovery failure. We have to evict the unreferenced + * lru inodes after clearing MS_ACTIVE because we don't + * otherwise clean up the lru if there's a subsequent failure in + * xfs_mountfs, which leads to us leaking the inodes if nothing + * else (e.g. quotacheck) references the inodes before the + * mount failure occurs. + */ + mp->m_super->s_flags |= MS_ACTIVE; error = xlog_recover_finish(mp->m_log); if (!error) xfs_log_work_queue(mp); + mp->m_super->s_flags &= ~MS_ACTIVE; + evict_inodes(mp->m_super); + + if (readonly) + mp->m_flags |= XFS_MOUNT_RDONLY; return error; } @@ -801,11 +831,14 @@ xfs_log_unmount_write(xfs_mount_t *mp) int error; /* - * Don't write out unmount record on read-only mounts. + * Don't write out unmount record on norecovery mounts or ro devices. * Or, if we are doing a forced umount (typically because of IO errors). */ - if (mp->m_flags & XFS_MOUNT_RDONLY) + if (mp->m_flags & XFS_MOUNT_NORECOVERY || + xfs_readonly_buftarg(log->l_mp->m_logdev_targp)) { + ASSERT(mp->m_flags & XFS_MOUNT_RDONLY); return 0; + } error = _xfs_log_force(mp, XFS_LOG_SYNC, NULL); ASSERT(error || !(XLOG_FORCED_SHUTDOWN(log))); @@ -825,9 +858,9 @@ xfs_log_unmount_write(xfs_mount_t *mp) if (!error) { /* the data section must be 32 bit size aligned */ struct { - __uint16_t magic; - __uint16_t pad1; - __uint32_t pad2; /* may as well make it 64 bits */ + uint16_t magic; + uint16_t pad1; + uint32_t pad2; /* may as well make it 64 bits */ } magic = { .magic = XLOG_UNMOUNT_TYPE, }; @@ -1189,8 +1222,7 @@ xlog_iodone(xfs_buf_t *bp) * IOABORT state. The IOABORT state is only set in DEBUG mode to inject * CRC errors into log recovery. */ - if (XFS_TEST_ERROR(bp->b_error, l->l_mp, XFS_ERRTAG_IODONE_IOERR, - XFS_RANDOM_IODONE_IOERR) || + if (XFS_TEST_ERROR(bp->b_error, l->l_mp, XFS_ERRTAG_IODONE_IOERR) || iclog->ic_state & XLOG_STATE_IOABORT) { if (iclog->ic_state & XLOG_STATE_IOABORT) iclog->ic_state &= ~XLOG_STATE_IOABORT; @@ -1665,7 +1697,7 @@ xlog_cksum( char *dp, int size) { - __uint32_t crc; + uint32_t crc; /* first generate the crc for the record header ... */ crc = xfs_start_cksum_update((char *)rhead, @@ -1828,7 +1860,7 @@ xlog_sync( */ dptr = (char *)&iclog->ic_header + count; for (i = 0; i < split; i += BBSIZE) { - __uint32_t cycle = be32_to_cpu(*(__be32 *)dptr); + uint32_t cycle = be32_to_cpu(*(__be32 *)dptr); if (++cycle == XLOG_HEADER_MAGIC_NUM) cycle++; *(__be32 *)dptr = cpu_to_be32(cycle); @@ -1842,7 +1874,6 @@ xlog_sync( /* calculcate the checksum */ iclog->ic_header.h_crc = xlog_cksum(log, &iclog->ic_header, iclog->ic_datap, size); -#ifdef DEBUG /* * Intentionally corrupt the log record CRC based on the error injection * frequency, if defined. This facilitates testing log recovery in the @@ -1850,15 +1881,13 @@ xlog_sync( * write on I/O completion and shutdown the fs. The subsequent mount * detects the bad CRC and attempts to recover. */ - if (log->l_badcrc_factor && - (prandom_u32() % log->l_badcrc_factor == 0)) { + if (XFS_TEST_ERROR(false, log->l_mp, XFS_ERRTAG_LOG_BAD_CRC)) { iclog->ic_header.h_crc &= cpu_to_le32(0xAAAAAAAA); iclog->ic_state |= XLOG_STATE_IOABORT; xfs_warn(log->l_mp, "Intentionally corrupted log record at LSN 0x%llx. Shutdown imminent.", be64_to_cpu(iclog->ic_header.h_lsn)); } -#endif bp->b_io_length = BTOBB(count); bp->b_fspriv = iclog; @@ -2024,7 +2053,7 @@ xlog_print_tic_res( }; #undef REG_TYPE_STR - xfs_warn(mp, "xlog_write: reservation summary:"); + xfs_warn(mp, "ticket reservation summary:"); xfs_warn(mp, " unit res = %d bytes", ticket->t_unit_res); xfs_warn(mp, " current res = %d bytes", @@ -2045,10 +2074,55 @@ xlog_print_tic_res( "bad-rtype" : res_type_str[r_type]), ticket->t_res_arr[i].r_len); } +} + +/* + * Print a summary of the transaction. + */ +void +xlog_print_trans( + struct xfs_trans *tp) +{ + struct xfs_mount *mp = tp->t_mountp; + struct xfs_log_item_desc *lidp; + + /* dump core transaction and ticket info */ + xfs_warn(mp, "transaction summary:"); + xfs_warn(mp, " flags = 0x%x", tp->t_flags); + + xlog_print_tic_res(mp, tp->t_ticket); + + /* dump each log item */ + list_for_each_entry(lidp, &tp->t_items, lid_trans) { + struct xfs_log_item *lip = lidp->lid_item; + struct xfs_log_vec *lv = lip->li_lv; + struct xfs_log_iovec *vec; + int i; + + xfs_warn(mp, "log item: "); + xfs_warn(mp, " type = 0x%x", lip->li_type); + xfs_warn(mp, " flags = 0x%x", lip->li_flags); + if (!lv) + continue; + xfs_warn(mp, " niovecs = %d", lv->lv_niovecs); + xfs_warn(mp, " size = %d", lv->lv_size); + xfs_warn(mp, " bytes = %d", lv->lv_bytes); + xfs_warn(mp, " buf len = %d", lv->lv_buf_len); + + /* dump each iovec for the log item */ + vec = lv->lv_iovecp; + for (i = 0; i < lv->lv_niovecs; i++) { + int dumplen = min(vec->i_len, 32); - xfs_alert_tag(mp, XFS_PTAG_LOGRES, - "xlog_write: reservation ran out. Need to up reservation"); - xfs_force_shutdown(mp, SHUTDOWN_LOG_IO_ERROR); + xfs_warn(mp, " iovec[%d]", i); + xfs_warn(mp, " type = 0x%x", vec->i_type); + xfs_warn(mp, " len = %d", vec->i_len); + xfs_warn(mp, " first %d bytes of iovec[%d]:", dumplen, i); + xfs_hex_dump(vec->i_addr, dumplen); + + vec++; + } + } } /* @@ -2321,8 +2395,12 @@ xlog_write( if (flags & (XLOG_COMMIT_TRANS | XLOG_UNMOUNT_TRANS)) ticket->t_curr_res -= sizeof(xlog_op_header_t); - if (ticket->t_curr_res < 0) + if (ticket->t_curr_res < 0) { + xfs_alert_tag(log->l_mp, XFS_PTAG_LOGRES, + "ctx ticket reservation ran out. Need to up reservation"); xlog_print_tic_res(log->l_mp, ticket); + xfs_force_shutdown(log->l_mp, SHUTDOWN_LOG_IO_ERROR); + } index = 0; lv = log_vector; @@ -2363,8 +2441,8 @@ xlog_write( } reg = &vecp[index]; - ASSERT(reg->i_len % sizeof(__int32_t) == 0); - ASSERT((unsigned long)ptr % sizeof(__int32_t) == 0); + ASSERT(reg->i_len % sizeof(int32_t) == 0); + ASSERT((unsigned long)ptr % sizeof(int32_t) == 0); start_rec_copy = xlog_write_start_rec(ptr, ticket); if (start_rec_copy) { @@ -2437,7 +2515,7 @@ next_lv: if (lv) vecp = lv->lv_iovecp; } - if (record_cnt == 0 && ordered == false) { + if (record_cnt == 0 && !ordered) { if (!lv) return 0; break; @@ -3143,7 +3221,7 @@ xlog_state_switch_iclogs( /* Round up to next log-sunit */ if (xfs_sb_version_haslogv2(&log->l_mp->m_sb) && log->l_mp->m_sb.sb_logsunit > 1) { - __uint32_t sunit_bb = BTOBB(log->l_mp->m_sb.sb_logsunit); + uint32_t sunit_bb = BTOBB(log->l_mp->m_sb.sb_logsunit); log->l_curr_block = roundup(log->l_curr_block, sunit_bb); } @@ -3297,8 +3375,6 @@ maybe_sleep: */ if (iclog->ic_state & XLOG_STATE_IOERROR) return -EIO; - if (log_flushed) - *log_flushed = 1; } else { no_sleep: @@ -3402,8 +3478,6 @@ try_again: xlog_wait(&iclog->ic_prev->ic_write_wait, &log->l_icloglock); - if (log_flushed) - *log_flushed = 1; already_slept = 1; goto try_again; } @@ -3437,9 +3511,6 @@ try_again: */ if (iclog->ic_state & XLOG_STATE_IOERROR) return -EIO; - - if (log_flushed) - *log_flushed = 1; } else { /* just return */ spin_unlock(&log->l_icloglock); } @@ -3771,7 +3842,7 @@ xlog_verify_iclog( xlog_in_core_2_t *xhdr; void *base_ptr, *ptr, *p; ptrdiff_t field_offset; - __uint8_t clientid; + uint8_t clientid; int len, i, j, k, op_len; int idx; |