From 9ca8dbcc65cfc63d6f5ef3312a33184e1d726e00 Mon Sep 17 00:00:00 2001 From: Yunhong Jiang Date: Tue, 4 Aug 2015 12:17:53 -0700 Subject: Add the rt linux 4.1.3-rt3 as base Import the rt linux 4.1.3-rt3 as OPNFV kvm base. It's from git://git.kernel.org/pub/scm/linux/kernel/git/rt/linux-rt-devel.git linux-4.1.y-rt and the base is: commit 0917f823c59692d751951bf5ea699a2d1e2f26a2 Author: Sebastian Andrzej Siewior Date: Sat Jul 25 12:13:34 2015 +0200 Prepare v4.1.3-rt3 Signed-off-by: Sebastian Andrzej Siewior We lose all the git history this way and it's not good. We should apply another opnfv project repo in future. Change-Id: I87543d81c9df70d99c5001fbdf646b202c19f423 Signed-off-by: Yunhong Jiang --- kernel/fs/dlm/recoverd.c | 342 +++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 342 insertions(+) create mode 100644 kernel/fs/dlm/recoverd.c (limited to 'kernel/fs/dlm/recoverd.c') diff --git a/kernel/fs/dlm/recoverd.c b/kernel/fs/dlm/recoverd.c new file mode 100644 index 000000000..6859b4bf9 --- /dev/null +++ b/kernel/fs/dlm/recoverd.c @@ -0,0 +1,342 @@ +/****************************************************************************** +******************************************************************************* +** +** Copyright (C) Sistina Software, Inc. 1997-2003 All rights reserved. +** Copyright (C) 2004-2011 Red Hat, Inc. All rights reserved. +** +** This copyrighted material is made available to anyone wishing to use, +** modify, copy, or redistribute it subject to the terms and conditions +** of the GNU General Public License v.2. +** +******************************************************************************* +******************************************************************************/ + +#include "dlm_internal.h" +#include "lockspace.h" +#include "member.h" +#include "dir.h" +#include "ast.h" +#include "recover.h" +#include "lowcomms.h" +#include "lock.h" +#include "requestqueue.h" +#include "recoverd.h" + + +/* If the start for which we're re-enabling locking (seq) has been superseded + by a newer stop (ls_recover_seq), we need to leave locking disabled. + + We suspend dlm_recv threads here to avoid the race where dlm_recv a) sees + locking stopped and b) adds a message to the requestqueue, but dlm_recoverd + enables locking and clears the requestqueue between a and b. */ + +static int enable_locking(struct dlm_ls *ls, uint64_t seq) +{ + int error = -EINTR; + + down_write(&ls->ls_recv_active); + + spin_lock(&ls->ls_recover_lock); + if (ls->ls_recover_seq == seq) { + set_bit(LSFL_RUNNING, &ls->ls_flags); + /* unblocks processes waiting to enter the dlm */ + up_write(&ls->ls_in_recovery); + clear_bit(LSFL_RECOVER_LOCK, &ls->ls_flags); + error = 0; + } + spin_unlock(&ls->ls_recover_lock); + + up_write(&ls->ls_recv_active); + return error; +} + +static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv) +{ + unsigned long start; + int error, neg = 0; + + log_rinfo(ls, "dlm_recover %llu", (unsigned long long)rv->seq); + + mutex_lock(&ls->ls_recoverd_active); + + dlm_callback_suspend(ls); + + dlm_clear_toss(ls); + + /* + * This list of root rsb's will be the basis of most of the recovery + * routines. + */ + + dlm_create_root_list(ls); + + /* + * Add or remove nodes from the lockspace's ls_nodes list. + */ + + error = dlm_recover_members(ls, rv, &neg); + if (error) { + log_rinfo(ls, "dlm_recover_members error %d", error); + goto fail; + } + + dlm_recover_dir_nodeid(ls); + + ls->ls_recover_dir_sent_res = 0; + ls->ls_recover_dir_sent_msg = 0; + ls->ls_recover_locks_in = 0; + + dlm_set_recover_status(ls, DLM_RS_NODES); + + error = dlm_recover_members_wait(ls); + if (error) { + log_rinfo(ls, "dlm_recover_members_wait error %d", error); + goto fail; + } + + start = jiffies; + + /* + * Rebuild our own share of the directory by collecting from all other + * nodes their master rsb names that hash to us. + */ + + error = dlm_recover_directory(ls); + if (error) { + log_rinfo(ls, "dlm_recover_directory error %d", error); + goto fail; + } + + dlm_set_recover_status(ls, DLM_RS_DIR); + + error = dlm_recover_directory_wait(ls); + if (error) { + log_rinfo(ls, "dlm_recover_directory_wait error %d", error); + goto fail; + } + + log_rinfo(ls, "dlm_recover_directory %u out %u messages", + ls->ls_recover_dir_sent_res, ls->ls_recover_dir_sent_msg); + + /* + * We may have outstanding operations that are waiting for a reply from + * a failed node. Mark these to be resent after recovery. Unlock and + * cancel ops can just be completed. + */ + + dlm_recover_waiters_pre(ls); + + error = dlm_recovery_stopped(ls); + if (error) + goto fail; + + if (neg || dlm_no_directory(ls)) { + /* + * Clear lkb's for departed nodes. + */ + + dlm_recover_purge(ls); + + /* + * Get new master nodeid's for rsb's that were mastered on + * departed nodes. + */ + + error = dlm_recover_masters(ls); + if (error) { + log_rinfo(ls, "dlm_recover_masters error %d", error); + goto fail; + } + + /* + * Send our locks on remastered rsb's to the new masters. + */ + + error = dlm_recover_locks(ls); + if (error) { + log_rinfo(ls, "dlm_recover_locks error %d", error); + goto fail; + } + + dlm_set_recover_status(ls, DLM_RS_LOCKS); + + error = dlm_recover_locks_wait(ls); + if (error) { + log_rinfo(ls, "dlm_recover_locks_wait error %d", error); + goto fail; + } + + log_rinfo(ls, "dlm_recover_locks %u in", + ls->ls_recover_locks_in); + + /* + * Finalize state in master rsb's now that all locks can be + * checked. This includes conversion resolution and lvb + * settings. + */ + + dlm_recover_rsbs(ls); + } else { + /* + * Other lockspace members may be going through the "neg" steps + * while also adding us to the lockspace, in which case they'll + * be doing the recover_locks (RS_LOCKS) barrier. + */ + dlm_set_recover_status(ls, DLM_RS_LOCKS); + + error = dlm_recover_locks_wait(ls); + if (error) { + log_rinfo(ls, "dlm_recover_locks_wait error %d", error); + goto fail; + } + } + + dlm_release_root_list(ls); + + /* + * Purge directory-related requests that are saved in requestqueue. + * All dir requests from before recovery are invalid now due to the dir + * rebuild and will be resent by the requesting nodes. + */ + + dlm_purge_requestqueue(ls); + + dlm_set_recover_status(ls, DLM_RS_DONE); + + error = dlm_recover_done_wait(ls); + if (error) { + log_rinfo(ls, "dlm_recover_done_wait error %d", error); + goto fail; + } + + dlm_clear_members_gone(ls); + + dlm_adjust_timeouts(ls); + + dlm_callback_resume(ls); + + error = enable_locking(ls, rv->seq); + if (error) { + log_rinfo(ls, "enable_locking error %d", error); + goto fail; + } + + error = dlm_process_requestqueue(ls); + if (error) { + log_rinfo(ls, "dlm_process_requestqueue error %d", error); + goto fail; + } + + error = dlm_recover_waiters_post(ls); + if (error) { + log_rinfo(ls, "dlm_recover_waiters_post error %d", error); + goto fail; + } + + dlm_recover_grant(ls); + + log_rinfo(ls, "dlm_recover %llu generation %u done: %u ms", + (unsigned long long)rv->seq, ls->ls_generation, + jiffies_to_msecs(jiffies - start)); + mutex_unlock(&ls->ls_recoverd_active); + + dlm_lsop_recover_done(ls); + return 0; + + fail: + dlm_release_root_list(ls); + log_rinfo(ls, "dlm_recover %llu error %d", + (unsigned long long)rv->seq, error); + mutex_unlock(&ls->ls_recoverd_active); + return error; +} + +/* The dlm_ls_start() that created the rv we take here may already have been + stopped via dlm_ls_stop(); in that case we need to leave the RECOVERY_STOP + flag set. */ + +static void do_ls_recovery(struct dlm_ls *ls) +{ + struct dlm_recover *rv = NULL; + + spin_lock(&ls->ls_recover_lock); + rv = ls->ls_recover_args; + ls->ls_recover_args = NULL; + if (rv && ls->ls_recover_seq == rv->seq) + clear_bit(LSFL_RECOVER_STOP, &ls->ls_flags); + spin_unlock(&ls->ls_recover_lock); + + if (rv) { + ls_recover(ls, rv); + kfree(rv->nodes); + kfree(rv); + } +} + +static int dlm_recoverd(void *arg) +{ + struct dlm_ls *ls; + + ls = dlm_find_lockspace_local(arg); + if (!ls) { + log_print("dlm_recoverd: no lockspace %p", arg); + return -1; + } + + down_write(&ls->ls_in_recovery); + set_bit(LSFL_RECOVER_LOCK, &ls->ls_flags); + wake_up(&ls->ls_recover_lock_wait); + + while (!kthread_should_stop()) { + set_current_state(TASK_INTERRUPTIBLE); + if (!test_bit(LSFL_RECOVER_WORK, &ls->ls_flags) && + !test_bit(LSFL_RECOVER_DOWN, &ls->ls_flags)) + schedule(); + set_current_state(TASK_RUNNING); + + if (test_and_clear_bit(LSFL_RECOVER_DOWN, &ls->ls_flags)) { + down_write(&ls->ls_in_recovery); + set_bit(LSFL_RECOVER_LOCK, &ls->ls_flags); + wake_up(&ls->ls_recover_lock_wait); + } + + if (test_and_clear_bit(LSFL_RECOVER_WORK, &ls->ls_flags)) + do_ls_recovery(ls); + } + + if (test_bit(LSFL_RECOVER_LOCK, &ls->ls_flags)) + up_write(&ls->ls_in_recovery); + + dlm_put_lockspace(ls); + return 0; +} + +int dlm_recoverd_start(struct dlm_ls *ls) +{ + struct task_struct *p; + int error = 0; + + p = kthread_run(dlm_recoverd, ls, "dlm_recoverd"); + if (IS_ERR(p)) + error = PTR_ERR(p); + else + ls->ls_recoverd_task = p; + return error; +} + +void dlm_recoverd_stop(struct dlm_ls *ls) +{ + kthread_stop(ls->ls_recoverd_task); +} + +void dlm_recoverd_suspend(struct dlm_ls *ls) +{ + wake_up(&ls->ls_wait_general); + mutex_lock(&ls->ls_recoverd_active); +} + +void dlm_recoverd_resume(struct dlm_ls *ls) +{ + mutex_unlock(&ls->ls_recoverd_active); +} + -- cgit 1.2.3-korg