/* -*- mode: c; c-basic-offset: 8; indent-tabs-mode: nil; -*- * vim:expandtab:shiftwidth=8:tabstop=8: * * lustre/mgc/mgc_request.c * Lustre Management Client config llog handling * * Copyright (C) 2006 Cluster File Systems, Inc. * Author Nathan Rutman * * This file is part of Lustre, http://www.lustre.org * * Lustre is free software; you can redistribute it and/or * modify it under the terms of version 2 of the GNU General Public * License as published by the Free Software Foundation. * * Lustre is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * * You should have received a copy of the GNU General Public License * along with Lustre; if not, write to the Free Software * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. * * For testing and management it is treated as an obd_device, * although * it does not export a full OBD method table (the * requests are coming * in over the wire, so object target modules * do not have a full * method table.) */ #ifndef EXPORT_SYMTAB # define EXPORT_SYMTAB #endif #define DEBUG_SUBSYSTEM S_MGC #define D_MGC D_CONFIG|D_WARNING #ifdef __KERNEL__ # include # include # include # include #else # include #endif #include #include #include #include #include int mgc_logname2resid(char *logname, struct ldlm_res_id *res_id) { char *name_end; int len; __u64 resname = 0; /* fsname is at most 8 chars long at the beginning of the logname e.g. "lustre-MDT0001" or "lustre" */ name_end = strrchr(logname, '-'); if (name_end) len = name_end - logname; else len = strlen(logname); if (len > 8) { CERROR("fsname too long: %s\n", logname); return -EINVAL; } memcpy(&resname, logname, len); memset(res_id, 0, sizeof(*res_id)); /* FIXME are resid names swabbed across the wire? */ res_id->name[0] = cpu_to_le64(resname); CDEBUG(D_MGC, "log %s to resid "LPX64"/"LPX64" (%.8s)\n", logname, res_id->name[0], res_id->name[1], (char *)&res_id->name[0]); return 0; } EXPORT_SYMBOL(mgc_logname2resid); /********************** config llog list **********************/ static struct list_head config_llog_list = LIST_HEAD_INIT(config_llog_list); static spinlock_t config_list_lock = SPIN_LOCK_UNLOCKED; static int config_log_get(struct config_llog_data *cld) { ENTRY; CDEBUG(D_INFO, "log %s refs %d\n", cld->cld_logname, atomic_read(&cld->cld_refcount)); atomic_inc(&cld->cld_refcount); if (cld->cld_stopping) { atomic_dec(&cld->cld_refcount); RETURN(1); } RETURN(0); } static void config_log_put(struct config_llog_data *cld) { ENTRY; CDEBUG(D_INFO, "log %s refs %d\n", cld->cld_logname, atomic_read(&cld->cld_refcount)); if (atomic_dec_and_test(&cld->cld_refcount)) { CDEBUG(D_MGC, "dropping config log %s\n", cld->cld_logname); OBD_FREE(cld->cld_logname, strlen(cld->cld_logname) + 1); if (cld->cld_cfg.cfg_instance != NULL) OBD_FREE(cld->cld_cfg.cfg_instance, strlen(cld->cld_cfg.cfg_instance) + 1); OBD_FREE(cld, sizeof(*cld)); } EXIT; } static struct config_llog_data *config_log_find(char *logname, struct config_llog_instance *cfg) { struct list_head *tmp; struct config_llog_data *cld; char *logid = logname; int match_instance = 0; ENTRY; if (cfg && cfg->cfg_instance) { match_instance++; logid = cfg->cfg_instance; } if (!logid) { CERROR("No log specified\n"); RETURN(ERR_PTR(-EINVAL)); } spin_lock(&config_list_lock); list_for_each(tmp, &config_llog_list) { cld = list_entry(tmp, struct config_llog_data, cld_list_chain); if (match_instance && cld->cld_cfg.cfg_instance && strcmp(logid, cld->cld_cfg.cfg_instance) == 0) goto out_found; if (!match_instance && strcmp(logid, cld->cld_logname) == 0) goto out_found; } spin_unlock(&config_list_lock); CERROR("can't get log %s\n", logid); RETURN(ERR_PTR(-ENOENT)); out_found: atomic_inc(&cld->cld_refcount); spin_unlock(&config_list_lock); RETURN(cld); } /* Add this log to our list of active logs. We have one active log per "mount" - client instance or servername. Each instance may be at a different point in the log. */ static int config_log_add(char *logname, struct config_llog_instance *cfg, struct super_block *sb) { struct config_llog_data *cld; int rc; ENTRY; CDEBUG(D_MGC, "adding config log %s:%s\n", logname, cfg->cfg_instance); OBD_ALLOC(cld, sizeof(*cld)); if (!cld) RETURN(-ENOMEM); OBD_ALLOC(cld->cld_logname, strlen(logname) + 1); if (!cld->cld_logname) { OBD_FREE(cld, sizeof(*cld)); RETURN(-ENOMEM); } strcpy(cld->cld_logname, logname); cld->cld_cfg = *cfg; cld->cld_cfg.cfg_last_idx = 0; cld->cld_cfg.cfg_flags = 0; cld->cld_cfg.cfg_sb = sb; atomic_set(&cld->cld_refcount, 1); if (cfg->cfg_instance != NULL) { OBD_ALLOC(cld->cld_cfg.cfg_instance, strlen(cfg->cfg_instance) + 1); strcpy(cld->cld_cfg.cfg_instance, cfg->cfg_instance); } rc = mgc_logname2resid(logname, &cld->cld_resid); if (rc) { config_log_put(cld); RETURN(rc); } spin_lock(&config_list_lock); list_add(&cld->cld_list_chain, &config_llog_list); spin_unlock(&config_list_lock); RETURN(rc); } /* Stop watching for updates on this log. */ static int config_log_end(char *logname, struct config_llog_instance *cfg) { struct config_llog_data *cld; int rc = 0; ENTRY; cld = config_log_find(logname, cfg); if (IS_ERR(cld)) RETURN(PTR_ERR(cld)); /* drop the ref from the find */ config_log_put(cld); cld->cld_stopping = 1; spin_lock(&config_list_lock); list_del(&cld->cld_list_chain); spin_unlock(&config_list_lock); /* drop the start ref */ config_log_put(cld); CDEBUG(D_MGC, "end config log %s (%d)\n", logname, rc); RETURN(rc); } /* Failsafe */ static void config_log_end_all(void) { struct list_head *tmp, *n; struct config_llog_data *cld; ENTRY; spin_lock(&config_list_lock); list_for_each_safe(tmp, n, &config_llog_list) { cld = list_entry(tmp, struct config_llog_data, cld_list_chain); CERROR("conflog failsafe %s\n", cld->cld_logname); list_del(&cld->cld_list_chain); config_log_put(cld); } spin_unlock(&config_list_lock); EXIT; } /********************** class fns **********************/ static int mgc_fs_setup(struct obd_device *obd, struct super_block *sb, struct vfsmount *mnt) { struct lvfs_run_ctxt saved; struct lustre_sb_info *lsi = s2lsi(sb); struct client_obd *cli = &obd->u.cli; struct dentry *dentry; int err = 0; ENTRY; LASSERT(lsi); LASSERT(lsi->lsi_srv_mnt == mnt); /* The mgc fs exclusion sem. Only one fs can be setup at a time. Maybe just overload the cl_sem? */ down(&cli->cl_mgc_sem); obd->obd_fsops = fsfilt_get_ops(MT_STR(lsi->lsi_ldd)); if (IS_ERR(obd->obd_fsops)) { up(&cli->cl_mgc_sem); CERROR("No fstype %s rc=%ld\n", MT_STR(lsi->lsi_ldd), PTR_ERR(obd->obd_fsops)); RETURN(PTR_ERR(obd->obd_fsops)); } cli->cl_mgc_vfsmnt = mnt; // FIXME which is the right SB? - filter_common_setup also CDEBUG(D_MGC, "SB's: fill=%p mnt=%p root=%p\n", sb, mnt->mnt_sb, mnt->mnt_root->d_inode->i_sb); fsfilt_setup(obd, mnt->mnt_sb); OBD_SET_CTXT_MAGIC(&obd->obd_lvfs_ctxt); obd->obd_lvfs_ctxt.pwdmnt = mnt; obd->obd_lvfs_ctxt.pwd = mnt->mnt_root; obd->obd_lvfs_ctxt.fs = get_ds(); push_ctxt(&saved, &obd->obd_lvfs_ctxt, NULL); dentry = lookup_one_len(MOUNT_CONFIGS_DIR, current->fs->pwd, strlen(MOUNT_CONFIGS_DIR)); pop_ctxt(&saved, &obd->obd_lvfs_ctxt, NULL); if (IS_ERR(dentry)) { err = PTR_ERR(dentry); CERROR("cannot lookup %s directory: rc = %d\n", MOUNT_CONFIGS_DIR, err); GOTO(err_ops, err); } cli->cl_mgc_configs_dir = dentry; /* We keep the cl_mgc_sem until mgc_fs_cleanup */ RETURN(0); err_ops: fsfilt_put_ops(obd->obd_fsops); obd->obd_fsops = NULL; cli->cl_mgc_vfsmnt = NULL; up(&cli->cl_mgc_sem); RETURN(err); } static int mgc_fs_cleanup(struct obd_device *obd) { struct client_obd *cli = &obd->u.cli; int rc = 0; ENTRY; LASSERT(cli->cl_mgc_vfsmnt != NULL); if (cli->cl_mgc_configs_dir != NULL) { struct lvfs_run_ctxt saved; push_ctxt(&saved, &obd->obd_lvfs_ctxt, NULL); l_dput(cli->cl_mgc_configs_dir); cli->cl_mgc_configs_dir = NULL; pop_ctxt(&saved, &obd->obd_lvfs_ctxt, NULL); } cli->cl_mgc_vfsmnt = NULL; if (obd->obd_fsops) fsfilt_put_ops(obd->obd_fsops); up(&cli->cl_mgc_sem); RETURN(rc); } static int mgc_cleanup(struct obd_device *obd) { struct client_obd *cli = &obd->u.cli; int rc; /* FIXME calls to mgc_fs_setup must take an obd ref to insure there's no fs by the time we get here. */ LASSERT(cli->cl_mgc_vfsmnt == NULL); rc = obd_llog_finish(obd, 0); if (rc != 0) CERROR("failed to cleanup llogging subsystems\n"); ptlrpcd_decref(); config_log_end_all(); return client_obd_cleanup(obd); } static struct obd_device *the_mgc; static int mgc_setup(struct obd_device *obd, obd_count len, void *buf) { int rc; ENTRY; ptlrpcd_addref(); rc = client_obd_setup(obd, len, buf); if (rc) GOTO(err_decref, rc); rc = obd_llog_init(obd, obd, 0, NULL); if (rc) { CERROR("failed to setup llogging subsystems\n"); GOTO(err_cleanup, rc); } the_mgc = obd; RETURN(rc); err_cleanup: client_obd_cleanup(obd); err_decref: ptlrpcd_decref(); RETURN(rc); } static int mgc_process_log(struct obd_device *mgc, struct config_llog_data *cld); /* FIXME I don't want a thread for every cld; make a list of cld's to requeue and use only 1 thread. */ /* reenqueue the lock, reparse the log */ static int mgc_async_requeue(void *data) { wait_queue_head_t waitq; struct l_wait_info lwi; struct config_llog_data *cld = (struct config_llog_data *)data; unsigned long flags; int rc = 0; ENTRY; if (!data) RETURN(-EINVAL); if (cld->cld_stopping) GOTO(out, rc = 0); lock_kernel(); ptlrpc_daemonize(); SIGNAL_MASK_LOCK(current, flags); sigfillset(¤t->blocked); RECALC_SIGPENDING; SIGNAL_MASK_UNLOCK(current, flags); THREAD_NAME(current->comm, sizeof(current->comm) - 1, "reQ %s", cld->cld_logname); unlock_kernel(); CDEBUG(D_MGC, "requeue "LPX64" %s:%s\n", cld->cld_resid.name[0], cld->cld_logname, cld->cld_cfg.cfg_instance); /* Sleep a few seconds to allow the server who caused the lock revocation to finish its setup, plus some random so everyone doesn't try to reconnect at once. */ init_waitqueue_head(&waitq); lwi = LWI_TIMEOUT(3 * HZ + (ll_rand() & 0x7f), NULL, NULL); l_wait_event(waitq, 0, &lwi); LASSERT(the_mgc); class_export_get(the_mgc->obd_self_export); #if 0 /* Re-send server info every time, in case MGS needs to regen its logs (for write_conf). Do we need this? It's extra RPCs for every server at every update. Turning it off until I'm sure it's needed. */ server_register_target(cld->cld_cfg.cfg_sb); #endif rc = mgc_process_log(the_mgc, cld); class_export_put(the_mgc->obd_self_export); out: /* Whether we enqueued again or not in mgc_process_log, we're done with the ref from the old mgc_blocking_ast */ config_log_put(cld); RETURN(rc); } /* based on ll_mdc_blocking_ast */ static int mgc_blocking_ast(struct ldlm_lock *lock, struct ldlm_lock_desc *desc, void *data, int flag) { struct lustre_handle lockh; struct config_llog_data *cld = (struct config_llog_data *)data; int rc = 0; ENTRY; switch (flag) { case LDLM_CB_BLOCKING: /* mgs wants the lock, give it up... */ LDLM_DEBUG(lock, "MGC blocking CB"); ldlm_lock2handle(lock, &lockh); rc = ldlm_cli_cancel(&lockh); break; case LDLM_CB_CANCELING: { /* We've given up the lock, prepare ourselves to update. */ LDLM_DEBUG(lock, "MGC cancel CB"); CDEBUG(D_MGC, "Lock res "LPX64" (%.8s)\n", lock->l_resource->lr_name.name[0], (char *)&lock->l_resource->lr_name.name[0]); /* Make sure not to re-enqueue when the mgc is stopping (we get called from client_disconnect_export) */ if (!lock->l_conn_export || !lock->l_conn_export->exp_obd->u.cli.cl_conn_count) { CDEBUG(D_MGC, "Disconnecting, don't requeue\n"); goto out_drop; } if (lock->l_req_mode != lock->l_granted_mode) { CERROR("original grant failed, won't requeue\n"); goto out_drop; } if (!data) { CERROR("missing data, won't requeue\n"); goto out_drop; } if (cld->cld_stopping) { CERROR("stopping, won't requeue\n"); goto out_drop; } /* Re-enqueue the lock in a separate thread, because we must return from this fn before that lock can be taken. */ rc = kernel_thread(mgc_async_requeue, data, CLONE_VM | CLONE_FS); if (rc < 0) { CERROR("Cannot re-enqueue thread: %d\n", rc); } else { rc = 0; break; } out_drop: /* Drop this here or in mgc_async_requeue, in either case, we're done with the reference after this. */ config_log_put(cld); break; } default: LBUG(); } if (rc) { CERROR("%s CB failed %d:\n", flag == LDLM_CB_BLOCKING ? "blocking" : "cancel", rc); LDLM_ERROR(lock, "MGC ast"); } RETURN(rc); } /* Take a config lock so we can get cancel notifications */ static int mgc_enqueue(struct obd_export *exp, struct lov_stripe_md *lsm, __u32 type, ldlm_policy_data_t *policy, __u32 mode, int *flags, void *bl_cb, void *cp_cb, void *gl_cb, void *data, __u32 lvb_len, void *lvb_swabber, struct lustre_handle *lockh) { struct config_llog_data *cld = (struct config_llog_data *)data; struct obd_device *obd = class_exp2obd(exp); int rc; ENTRY; CDEBUG(D_MGC, "Enqueue for %s (res "LPX64")\n", cld->cld_logname, cld->cld_resid.name[0]); /* We can only drop this config log ref when we drop the lock */ if (config_log_get(cld)) RETURN(ELDLM_LOCK_ABORTED); /* We need a callback for every lockholder, so don't try to ldlm_lock_match (see rev 1.1.2.11.2.47) */ rc = ldlm_cli_enqueue(exp, NULL, obd->obd_namespace, cld->cld_resid, type, NULL, mode, flags, mgc_blocking_ast, ldlm_completion_ast, NULL, data, NULL, 0, NULL, lockh); RETURN(rc); } static int mgc_cancel(struct obd_export *exp, struct lov_stripe_md *md, __u32 mode, struct lustre_handle *lockh) { ENTRY; ldlm_lock_decref(lockh, mode); RETURN(0); } #if 0 static int mgc_iocontrol(unsigned int cmd, struct obd_export *exp, int len, void *karg, void *uarg) { struct obd_device *obd = exp->exp_obd; struct obd_ioctl_data *data = karg; struct llog_ctxt *ctxt; struct lvfs_run_ctxt saved; int rc; ENTRY; #if (LINUX_VERSION_CODE < KERNEL_VERSION(2,5,0)) MOD_INC_USE_COUNT; #else if (!try_module_get(THIS_MODULE)) { CERROR("Can't get module. Is it alive?"); return -EINVAL; } #endif switch (cmd) { /* REPLicator context */ case OBD_IOC_PARSE: { CERROR("MGC parsing llog %s\n", data->ioc_inlbuf1); ctxt = llog_get_context(exp->exp_obd, LLOG_CONFIG_REPL_CTXT); rc = class_config_parse_llog(ctxt, data->ioc_inlbuf1, NULL); GOTO(out, rc); } #ifdef __KERNEL__ case OBD_IOC_LLOG_INFO: case OBD_IOC_LLOG_PRINT: { ctxt = llog_get_context(obd, LLOG_CONFIG_REPL_CTXT); rc = llog_ioctl(ctxt, cmd, data); GOTO(out, rc); } #endif /* ORIGinator context */ case OBD_IOC_DUMP_LOG: { ctxt = llog_get_context(obd, LLOG_CONFIG_ORIG_CTXT); push_ctxt(&saved, &obd->obd_lvfs_ctxt, NULL); rc = class_config_dump_llog(ctxt, data->ioc_inlbuf1, NULL); pop_ctxt(&saved, &obd->obd_lvfs_ctxt, NULL); if (rc) RETURN(rc); GOTO(out, rc); } default: CERROR("mgc_ioctl(): unrecognised ioctl %#x\n", cmd); GOTO(out, rc = -ENOTTY); } out: #if (LINUX_VERSION_CODE < KERNEL_VERSION(2,5,0)) MOD_DEC_USE_COUNT; #else module_put(THIS_MODULE); #endif return rc; } #endif /* Send target_reg message to MGS */ static int mgc_target_register(struct obd_export *exp, struct mgs_target_info *mti) { struct ptlrpc_request *req; struct mgs_target_info *req_mti, *rep_mti; int size = sizeof(*req_mti); int rep_size = sizeof(*mti); int rc; ENTRY; req = ptlrpc_prep_req(class_exp2cliimp(exp), LUSTRE_MGS_VERSION, MGS_TARGET_REG, 1, &size, NULL); if (!req) RETURN(-ENOMEM); req_mti = lustre_msg_buf(req->rq_reqmsg, 0, sizeof(*req_mti)); if (!req_mti) RETURN(-ENOMEM); memcpy(req_mti, mti, sizeof(*req_mti)); req->rq_replen = lustre_msg_size(1, &rep_size); CDEBUG(D_MGC, "register %s\n", mti->mti_svname); rc = ptlrpc_queue_wait(req); if (!rc) { rep_mti = lustre_swab_repbuf(req, 0, sizeof(*rep_mti), lustre_swab_mgs_target_info); memcpy(mti, rep_mti, sizeof(*rep_mti)); CDEBUG(D_MGC, "register %s got index = %d\n", mti->mti_svname, mti->mti_stripe_index); } else { CERROR("register failed. rc=%d\n", rc); } ptlrpc_req_finished(req); RETURN(rc); } int mgc_set_info(struct obd_export *exp, obd_count keylen, void *key, obd_count vallen, void *val) { struct obd_import *imp = class_exp2cliimp(exp); int rc = -EINVAL; ENTRY; /* Try to "recover" the initial connection; i.e. retry */ if (KEY_IS(KEY_INIT_RECOV)) { if (vallen != sizeof(int)) RETURN(-EINVAL); imp->imp_initial_recov = *(int *)val; CDEBUG(D_HA, "%s: set imp_initial_recov = %d\n", exp->exp_obd->obd_name, imp->imp_initial_recov); RETURN(0); } /* Turn off initial_recov after we try all backup servers once */ if (KEY_IS(KEY_INIT_RECOV_BACKUP)) { if (vallen != sizeof(int)) RETURN(-EINVAL); imp->imp_initial_recov_bk = *(int *)val; CDEBUG(D_HA, "%s: set imp_initial_recov_bk = %d\n", exp->exp_obd->obd_name, imp->imp_initial_recov_bk); if (imp->imp_invalid) { /* Resurrect if we previously died */ CDEBUG(D_MGC, "Reactivate %s %d:%d:%d\n", imp->imp_obd->obd_name, imp->imp_deactive, imp->imp_invalid, imp->imp_state); /* can't put this in obdclass, module loop with ptlrpc*/ /* This seems to be necessary when restarting a combo mgs/mdt while the mgc is alive */ ptlrpc_invalidate_import(imp); /* Remove 'invalid' flag */ ptlrpc_activate_import(imp); /* Attempt a new connect */ ptlrpc_recover_import(imp, NULL); } RETURN(0); } /* Hack alert */ if (KEY_IS("register_target")) { struct mgs_target_info *mti; if (vallen != sizeof(struct mgs_target_info)) RETURN(-EINVAL); mti = (struct mgs_target_info *)val; CDEBUG(D_MGC, "register_target %s %#x\n", mti->mti_svname, mti->mti_flags); rc = mgc_target_register(exp, mti); RETURN(rc); } if (KEY_IS("set_fs")) { struct super_block *sb = (struct super_block *)val; struct lustre_sb_info *lsi; if (vallen != sizeof(struct super_block)) RETURN(-EINVAL); lsi = s2lsi(sb); rc = mgc_fs_setup(exp->exp_obd, sb, lsi->lsi_srv_mnt); if (rc) { CERROR("set_fs got %d\n", rc); } RETURN(rc); } if (KEY_IS("clear_fs")) { if (vallen != 0) RETURN(-EINVAL); rc = mgc_fs_cleanup(exp->exp_obd); if (rc) { CERROR("clear_fs got %d\n", rc); } RETURN(rc); } RETURN(rc); } static int mgc_import_event(struct obd_device *obd, struct obd_import *imp, enum obd_import_event event) { int rc = 0; LASSERT(imp->imp_obd == obd); CDEBUG(D_MGC, "import event %#x\n", event); switch (event) { case IMP_EVENT_INVALIDATE: { struct ldlm_namespace *ns = obd->obd_namespace; ldlm_namespace_cleanup(ns, LDLM_FL_LOCAL_ONLY); break; } case IMP_EVENT_DISCON: case IMP_EVENT_INACTIVE: case IMP_EVENT_ACTIVE: case IMP_EVENT_OCD: break; default: CERROR("Unknown import event %#x\n", event); LBUG(); } RETURN(rc); } static int mgc_llog_init(struct obd_device *obd, struct obd_device *tgt, int count, struct llog_catid *logid) { struct llog_ctxt *ctxt; int rc; ENTRY; rc = llog_setup(obd, LLOG_CONFIG_ORIG_CTXT, tgt, 0, NULL, &llog_lvfs_ops); if (rc) RETURN(rc); rc = llog_setup(obd, LLOG_CONFIG_REPL_CTXT, tgt, 0, NULL, &llog_client_ops); if (rc == 0) { ctxt = llog_get_context(obd, LLOG_CONFIG_REPL_CTXT); ctxt->loc_imp = obd->u.cli.cl_import; } RETURN(rc); } static int mgc_llog_finish(struct obd_device *obd, int count) { int rc; ENTRY; rc = llog_cleanup(llog_get_context(obd, LLOG_CONFIG_REPL_CTXT)); rc = llog_cleanup(llog_get_context(obd, LLOG_CONFIG_ORIG_CTXT)); RETURN(rc); } /* identical to mgs_log_is_empty */ static int mgc_llog_is_empty(struct obd_device *obd, struct llog_ctxt *ctxt, char *name) { struct lvfs_run_ctxt saved; struct llog_handle *llh; int rc = 0; push_ctxt(&saved, &obd->obd_lvfs_ctxt, NULL); rc = llog_create(ctxt, &llh, NULL, name); if (rc == 0) { llog_init_handle(llh, LLOG_F_IS_PLAIN, NULL); rc = llog_get_size(llh); llog_close(llh); } pop_ctxt(&saved, &obd->obd_lvfs_ctxt, NULL); /* header is record 1 */ return(rc <= 1); } static int mgc_copy_handler(struct llog_handle *llh, struct llog_rec_hdr *rec, void *data) { struct llog_rec_hdr local_rec = *rec; struct llog_handle *local_llh = (struct llog_handle *)data; char *cfg_buf = (char*) (rec + 1); struct lustre_cfg *lcfg; int rc = 0; ENTRY; lcfg = (struct lustre_cfg *)cfg_buf; /* FIXME we should always write to an empty log, so remove this check.*/ /* append new records */ if (rec->lrh_index >= llog_get_size(local_llh)) { rc = llog_write_rec(local_llh, &local_rec, NULL, 0, (void *)cfg_buf, -1); CDEBUG(D_INFO, "idx=%d, rc=%d, len=%d, cmd %x %s %s\n", rec->lrh_index, rc, rec->lrh_len, lcfg->lcfg_command, lustre_cfg_string(lcfg, 0), lustre_cfg_string(lcfg, 1)); } else { CDEBUG(D_INFO, "skip idx=%d\n", rec->lrh_index); } RETURN(rc); } static int mgc_copy_llog(struct obd_device *obd, struct llog_ctxt *rctxt, struct llog_ctxt *lctxt, char *logname) { struct llog_handle *local_llh, *remote_llh; struct obd_uuid *uuid; int rc, rc2; ENTRY; /* open local log */ rc = llog_create(lctxt, &local_llh, NULL, logname); if (rc) RETURN(rc); /* set the log header uuid for fun */ OBD_ALLOC_PTR(uuid); obd_str2uuid(uuid, logname); rc = llog_init_handle(local_llh, LLOG_F_IS_PLAIN, uuid); OBD_FREE_PTR(uuid); if (rc) GOTO(out_closel, rc); /* FIXME write new log to a temp name, then vfs_rename over logname upon successful completion. */ /* open remote log */ rc = llog_create(rctxt, &remote_llh, NULL, logname); if (rc) GOTO(out_closel, rc); rc = llog_init_handle(remote_llh, LLOG_F_IS_PLAIN, NULL); if (rc) GOTO(out_closer, rc); rc = llog_process(remote_llh, mgc_copy_handler,(void *)local_llh, NULL); out_closer: rc2 = llog_close(remote_llh); if (!rc) rc = rc2; out_closel: rc2 = llog_close(local_llh); if (!rc) rc = rc2; CDEBUG(D_MGC, "Copied remote log %s (%d)\n", logname, rc); RETURN(rc); } DECLARE_MUTEX(llog_process_lock); /* Get a config log from the MGS and process it. This func is called for both clients and servers. */ static int mgc_process_log(struct obd_device *mgc, struct config_llog_data *cld) { struct llog_ctxt *ctxt, *lctxt; struct lustre_handle lockh; struct client_obd *cli = &mgc->u.cli; struct lvfs_run_ctxt saved; struct lustre_sb_info *lsi; int rc, rcl, flags = 0, must_pop = 0; ENTRY; if (!cld || !cld->cld_cfg.cfg_sb) { /* This should never happen */ CERROR("Missing cld, aborting log update\n"); RETURN(-EINVAL); } if (cld->cld_stopping) RETURN(0); lsi = s2lsi(cld->cld_cfg.cfg_sb); CDEBUG(D_MGC, "Process log %s:%s from %d\n", cld->cld_logname, cld->cld_cfg.cfg_instance, cld->cld_cfg.cfg_last_idx + 1); ctxt = llog_get_context(mgc, LLOG_CONFIG_REPL_CTXT); if (!ctxt) { CERROR("missing llog context\n"); RETURN(-EINVAL); } /* I don't want mutliple processes running process_log at once -- sounds like badness. It actually might be fine, as long as we're not trying to update from the same log simultaneously (in which case we should use a per-log sem.) */ down(&llog_process_lock); /* Get the cfg lock on the llog */ rcl = mgc_enqueue(mgc->u.cli.cl_mgc_mgsexp, NULL, LDLM_PLAIN, NULL, LCK_CR, &flags, NULL, NULL, NULL, cld, 0, NULL, &lockh); if (rcl) CERROR("Can't get cfg lock: %d\n", rcl); lctxt = llog_get_context(mgc, LLOG_CONFIG_ORIG_CTXT); /* Copy the setup log locally if we can. Don't mess around if we're running an MGS though (logs are already local). */ if (lctxt && lsi && (lsi->lsi_flags & LSI_SERVER) && (lsi->lsi_srv_mnt == cli->cl_mgc_vfsmnt) && !IS_MGS(lsi->lsi_ldd)) { push_ctxt(&saved, &mgc->obd_lvfs_ctxt, NULL); must_pop++; if (rcl == 0) /* Only try to copy log if we have the lock. */ rc = mgc_copy_llog(mgc, ctxt, lctxt, cld->cld_logname); if (rcl || rc) { if (mgc_llog_is_empty(mgc, lctxt, cld->cld_logname)) { LCONSOLE_ERROR("Failed to get MGS log %s " "and no local copy.\n", cld->cld_logname); GOTO(out_pop, rc = -ENOTCONN); } LCONSOLE_WARN("Failed to get MGS log %s, using " "local copy.\n", cld->cld_logname); } /* Now, whether we copied or not, start using the local llog. If we failed to copy, we'll start using whatever the old log has. */ ctxt = lctxt; } /* logname and instance info should be the same, so use our copy of the instance for the update. The cfg_last_idx will be updated here. */ rc = class_config_parse_llog(ctxt, cld->cld_logname, &cld->cld_cfg); out_pop: if (must_pop) pop_ctxt(&saved, &mgc->obd_lvfs_ctxt, NULL); /* Now drop the lock so MGS can revoke it */ if (!rcl) { rcl = mgc_cancel(mgc->u.cli.cl_mgc_mgsexp, NULL, LCK_CR, &lockh); if (rcl) CERROR("Can't drop cfg lock: %d\n", rcl); } if (rc) { CERROR("%s: the configuration '%s' could not be read " "(%d) from the MGS.\n", mgc->obd_name, cld->cld_logname, rc); } up(&llog_process_lock); RETURN(rc); } static int mgc_process_config(struct obd_device *obd, obd_count len, void *buf) { struct lustre_cfg *lcfg = buf; int cmd; int rc = 0; ENTRY; switch(cmd = lcfg->lcfg_command) { case LCFG_LOV_ADD_OBD: { struct mgs_target_info *mti; if (LUSTRE_CFG_BUFLEN(lcfg, 1) != sizeof(struct mgs_target_info)) GOTO(out, rc = -EINVAL); mti = (struct mgs_target_info *)lustre_cfg_buf(lcfg, 1); CDEBUG(D_MGC, "add_target %s %#x\n", mti->mti_svname, mti->mti_flags); rc = mgc_target_register(obd->u.cli.cl_mgc_mgsexp, mti); break; } case LCFG_LOV_DEL_OBD: /* FIXME */ CERROR("lov_del_obd unimplemented\n"); rc = -ENOSYS; break; case LCFG_LOG_START: { struct config_llog_data *cld; struct config_llog_instance *cfg; struct super_block *sb; char *logname = lustre_cfg_string(lcfg, 1); cfg = (struct config_llog_instance *)lustre_cfg_buf(lcfg, 2); sb = *(struct super_block **)lustre_cfg_buf(lcfg, 3); CDEBUG(D_MGC, "parse_log %s from %d\n", logname, cfg->cfg_last_idx); /* We're only called through here on the initial mount */ rc = config_log_add(logname, cfg, sb); if (rc) break; cld = config_log_find(logname, cfg); if (IS_ERR(cld)) { rc = PTR_ERR(cld); break; } /* COMPAT_146 */ /* For old logs, there was no start marker. */ /* FIXME only set this for old logs! */ cld->cld_cfg.cfg_flags |= CFG_F_MARKER; rc = mgc_process_log(obd, cld); config_log_put(cld); break; } case LCFG_LOG_END: { struct config_llog_instance *cfg = NULL; char *logname = lustre_cfg_string(lcfg, 1); if (lcfg->lcfg_bufcount >= 2) cfg = (struct config_llog_instance *)lustre_cfg_buf( lcfg, 2); rc = config_log_end(logname, cfg); break; } default: { CERROR("Unknown command: %d\n", lcfg->lcfg_command); GOTO(out, rc = -EINVAL); } } out: RETURN(rc); } struct obd_ops mgc_obd_ops = { .o_owner = THIS_MODULE, .o_setup = mgc_setup, .o_cleanup = mgc_cleanup, .o_add_conn = client_import_add_conn, .o_del_conn = client_import_del_conn, .o_connect = client_connect_import, .o_disconnect = client_disconnect_export, //.o_enqueue = mgc_enqueue, .o_cancel = mgc_cancel, //.o_iocontrol = mgc_iocontrol, .o_set_info = mgc_set_info, .o_import_event = mgc_import_event, .o_llog_init = mgc_llog_init, .o_llog_finish = mgc_llog_finish, .o_process_config = mgc_process_config, }; int __init mgc_init(void) { return class_register_type(&mgc_obd_ops, NULL, LUSTRE_MGC_NAME); } #ifdef __KERNEL__ static void /*__exit*/ mgc_exit(void) { class_unregister_type(LUSTRE_MGC_NAME); } MODULE_AUTHOR("Cluster File Systems, Inc. "); MODULE_DESCRIPTION("Lustre Management Client"); MODULE_LICENSE("GPL"); module_init(mgc_init); module_exit(mgc_exit); #endif