/*- * See the file LICENSE for redistribution information. * * Copyright (c) 1996, 1997 * Sleepycat Software. All rights reserved. */ #include "config.h" #ifndef lint static const char sccsid[] = "@(#)log.c 10.27 (Sleepycat) 9/23/97"; #endif /* not lint */ #ifndef NO_SYSTEM_INCLUDES #include #include #include #include #include #include #include #endif #include "db_int.h" #include "shqueue.h" #include "db_shash.h" #include "log.h" #include "db_dispatch.h" #include "txn_auto.h" #include "common_ext.h" static int __log_recover __P((DB_LOG *)); /* * log_open -- * Initialize and/or join a log. */ int log_open(path, flags, mode, dbenv, lpp) const char *path; int flags; int mode; DB_ENV *dbenv; DB_LOG **lpp; { DB_LOG *dblp; LOG *lp; size_t len; int fd, newregion, ret, retry_cnt; /* Validate arguments. */ #ifdef HAVE_SPINLOCKS #define OKFLAGS (DB_CREATE | DB_THREAD) #else #define OKFLAGS (DB_CREATE) #endif if ((ret = __db_fchk(dbenv, "log_open", flags, OKFLAGS)) != 0) return (ret); /* * We store 4-byte offsets into the file, so the maximum file * size can't be larger than that. */ if (dbenv != NULL && dbenv->lg_max > UINT32_T_MAX) { __db_err(dbenv, "log_open: maximum file size too large"); return (EINVAL); } /* Create and initialize the DB_LOG structure. */ if ((dblp = (DB_LOG *)calloc(1, sizeof(DB_LOG))) == NULL) return (ENOMEM); if (path != NULL && (dblp->dir = strdup(path)) == NULL) { free(dblp); return (ENOMEM); } dblp->dbenv = dbenv; dblp->lfd = -1; ZERO_LSN(dblp->c_lsn); dblp->c_fd = -1; /* * The log region isn't fixed size because we store the registered * file names there. Make it fairly large so that we don't have to * grow it. */ len = 30 * 1024; /* Map in the region. */ retry_cnt = newregion = 0; retry: if (LF_ISSET(DB_CREATE)) { ret = __db_rcreate(dbenv, DB_APP_LOG, path, DB_DEFAULT_LOG_FILE, mode, len, &fd, &dblp->maddr); if (ret == 0) { /* Put the LOG structure first in the region. */ lp = dblp->maddr; /* Initialize the rest of the region as free space. */ dblp->addr = (u_int8_t *)dblp->maddr + sizeof(LOG); __db_shalloc_init(dblp->addr, len - sizeof(LOG)); /* Initialize the LOG structure. */ lp->persist.lg_max = dbenv == NULL ? 0 : dbenv->lg_max; if (lp->persist.lg_max == 0) lp->persist.lg_max = DEFAULT_MAX; lp->persist.magic = DB_LOGMAGIC; lp->persist.version = DB_LOGVERSION; lp->persist.mode = mode; SH_TAILQ_INIT(&lp->fq); /* Initialize LOG LSNs. */ lp->lsn.file = 1; lp->lsn.offset = 0; newregion = 1; } else if (ret != EEXIST) goto err; } /* If we didn't or couldn't create the region, try and join it. */ if (!newregion && (ret = __db_ropen(dbenv, DB_APP_LOG, path, DB_DEFAULT_LOG_FILE, 0, &fd, &dblp->maddr)) != 0) { /* * If we fail because the file isn't available, wait a * second and try again. */ if (ret == EAGAIN && ++retry_cnt < 3) { (void)__db_sleep(1, 0); goto retry; } goto err; } /* Set up the common information. */ dblp->lp = dblp->maddr; dblp->addr = (u_int8_t *)dblp->maddr + sizeof(LOG); dblp->fd = fd; /* Initialize thread information. */ if (LF_ISSET(DB_THREAD)) { F_SET(dblp, DB_AM_THREAD); if (!newregion) LOCK_LOGREGION(dblp); if ((ret = __db_shalloc(dblp->addr, sizeof(db_mutex_t), MUTEX_ALIGNMENT, &dblp->mutexp)) == 0) (void)__db_mutex_init(dblp->mutexp, -1); if (!newregion) UNLOCK_LOGREGION(dblp); if (ret != 0) { (void)log_close(dblp); if (newregion) (void)log_unlink(path, 1, dbenv); return (ret); } } /* * If doing recovery, try and recover any previous log files * before releasing the lock. */ if (newregion) { ret = __log_recover(dblp); UNLOCK_LOGREGION(dblp); if (ret != 0) { (void)log_close(dblp); (void)log_unlink(path, 1, dbenv); return (ret); } } *lpp = dblp; return (0); err: /* * We never get here with an allocated thread-mutex, so we do * not have to worry about freeing it. */ FREE(dblp, sizeof(DB_LOG)); return (ret); } /* * __log_recover -- * Recover a log. */ static int __log_recover(dblp) DB_LOG *dblp; { DBT dbt; DB_LSN lsn; LOG *lp; u_int32_t chk; int cnt, found_checkpoint, ret; lp = dblp->lp; /* * Find a log file. If none exist, we simply return, leaving * everything initialized to a new log. */ if ((ret = __log_find(dblp, &cnt)) != 0) return (ret); if (cnt == 0) return (0); /* We have a log file name, find the last one. */ while (cnt < MAXLFNAME) if (__log_valid(dblp, lp, ++cnt) != 0) { --cnt; break; } /* * We have the last useful log file and we've loaded any persistent * information. Pretend that the log is larger than it can possibly * be, and read this file, looking for a checkpoint and its end. */ dblp->c_lsn.file = cnt; dblp->c_lsn.offset = 0; lsn = dblp->c_lsn; lp->lsn.file = cnt + 1; lp->lsn.offset = 0; /* Set the cursor. Shouldn't fail, leave error messages on. */ memset(&dbt, 0, sizeof(dbt)); if ((ret = __log_get(dblp, &lsn, &dbt, DB_SET, 0)) != 0) return (ret); /* * Read to the end of the file, saving checkpoints. This will fail * at some point, so turn off error messages. */ found_checkpoint = 0; while (__log_get(dblp, &lsn, &dbt, DB_NEXT, 1) == 0) { if (dbt.size < sizeof(u_int32_t)) continue; memcpy(&chk, dbt.data, sizeof(u_int32_t)); if (chk == DB_txn_ckp) { lp->c_lsn = lsn; found_checkpoint = 1; } } /* * We know where the end of the log is. Since that record is on disk, * it's also the last-synced LSN. */ lp->lsn = lsn; lp->lsn.offset += dblp->c_len; lp->s_lsn = lp->lsn; /* Set up the current buffer information, too. */ lp->len = dblp->c_len; lp->b_off = 0; lp->w_off = lp->lsn.offset; /* * It's possible that we didn't find a checkpoint because there wasn't * one in the last log file. Start searching. */ while (!found_checkpoint && cnt > 1) { dblp->c_lsn.file = --cnt; dblp->c_lsn.offset = 0; lsn = dblp->c_lsn; /* Set the cursor. Shouldn't fail, leave error messages on. */ if ((ret = __log_get(dblp, &lsn, &dbt, DB_SET, 0)) != 0) return (ret); /* * Read to the end of the file, saving checkpoints. Shouldn't * fail, leave error messages on. */ while (__log_get(dblp, &lsn, &dbt, DB_NEXT, 0) == 0) { if (dbt.size < sizeof(u_int32_t)) continue; memcpy(&chk, dbt.data, sizeof(u_int32_t)); if (chk == DB_txn_ckp) { lp->c_lsn = lsn; found_checkpoint = 1; } } } /* If we never find a checkpoint, that's okay, just 0 it out. */ if (!found_checkpoint) { lp->c_lsn.file = 1; lp->c_lsn.offset = 0; } __db_err(dblp->dbenv, "Recovering the log: last valid LSN: file: %lu offset %lu", (u_long)lp->lsn.file, (u_long)lp->lsn.offset); /* Reset the cursor. */ ZERO_LSN(dblp->c_lsn); return (0); } /* * __log_find -- * Try to find a log file. * * PUBLIC: int __log_find __P((DB_LOG *, int *)); */ int __log_find(dblp, valp) DB_LOG *dblp; int *valp; { int cnt, fcnt, logval, ret; const char *dir; char **names, *p, *q; /* Find the directory name. */ if ((ret = __log_name(dblp, 1, &p)) != 0) return (ret); if ((q = __db_rpath(p)) == NULL) dir = PATH_DOT; else { *q = '\0'; dir = p; } /* Get the list of file names. */ ret = __db_dir(dblp->dbenv, dir, &names, &fcnt); FREES(p); if (ret != 0) return (ret); /* * Search for a valid log file name, return a value of 0 on * failure. */ *valp = 0; for (cnt = fcnt, logval = 0; --cnt >= 0;) if (strncmp(names[cnt], "log.", sizeof("log.") - 1) == 0) { logval = atoi(names[cnt] + 4); if (logval != 0 && __log_valid(dblp, dblp->lp, logval) == 0) { *valp = logval; break; } } /* Discard the list. */ __db_dirf(dblp->dbenv, names, fcnt); return (ret); } /* * log_valid -- * Validate a log file. * * PUBLIC: int __log_valid __P((DB_LOG *, LOG *, int)); */ int __log_valid(dblp, lp, cnt) DB_LOG *dblp; LOG *lp; int cnt; { LOGP persist; ssize_t nw; int fd, ret; char *p; if ((ret = __log_name(dblp, cnt, &p)) != 0) return (ret); fd = -1; if ((ret = __db_fdopen(p, DB_RDONLY | DB_SEQUENTIAL, DB_RDONLY | DB_SEQUENTIAL, 0, &fd)) != 0 || (ret = __db_lseek(fd, 0, 0, sizeof(HDR), SEEK_SET)) != 0 || (ret = __db_read(fd, &persist, sizeof(LOGP), &nw)) != 0 || nw != sizeof(LOGP)) { if (ret == 0) ret = EIO; if (fd != -1) { (void)__db_close(fd); __db_err(dblp->dbenv, "Ignoring log file: %s: %s", p, strerror(ret)); } goto err; } (void)__db_close(fd); if (persist.magic != DB_LOGMAGIC) { __db_err(dblp->dbenv, "Ignoring log file: %s: magic number %lx, not %lx", p, (u_long)persist.magic, (u_long)DB_LOGMAGIC); ret = EINVAL; goto err; } if (persist.version < DB_LOGOLDVER || persist.version > DB_LOGVERSION) { __db_err(dblp->dbenv, "Ignoring log file: %s: unsupported log version %lu", p, (u_long)persist.version); ret = EINVAL; goto err; } if (lp != NULL) { lp->persist.lg_max = persist.lg_max; lp->persist.mode = persist.mode; } ret = 0; err: FREES(p); return (ret); } /* * log_close -- * Close a log. */ int log_close(dblp) DB_LOG *dblp; { int ret, t_ret; ret = 0; /* Discard the per-thread pointer. */ if (dblp->mutexp != NULL) { LOCK_LOGREGION(dblp); __db_shalloc_free(dblp->addr, dblp->mutexp); UNLOCK_LOGREGION(dblp); } /* Close the region. */ if ((t_ret = __db_rclose(dblp->dbenv, dblp->fd, dblp->maddr)) != 0 && ret == 0) ret = t_ret; /* Close open files, release allocated memory. */ if (dblp->lfd != -1 && (t_ret = __db_close(dblp->lfd)) != 0 && ret == 0) ret = t_ret; if (dblp->c_dbt.data != NULL) FREE(dblp->c_dbt.data, dblp->c_dbt.ulen); if (dblp->c_fd != -1 && (t_ret = __db_close(dblp->c_fd)) != 0 && ret == 0) ret = t_ret; if (dblp->dbentry != NULL) FREE(dblp->dbentry, (dblp->dbentry_cnt * sizeof(DB_ENTRY))); if (dblp->dir != NULL) FREES(dblp->dir); /* Free the structure. */ FREE(dblp, sizeof(DB_LOG)); return (ret); } /* * log_unlink -- * Exit a log. */ int log_unlink(path, force, dbenv) const char *path; int force; DB_ENV *dbenv; { return (__db_runlink(dbenv, DB_APP_LOG, path, DB_DEFAULT_LOG_FILE, force)); }