/* -*- mode: c; c-basic-offset: 8; indent-tabs-mode: nil; -*-
- * (visit-tags-table FILE)
* vim:expandtab:shiftwidth=8:tabstop=8:
+ *
+ * GPL HEADER START
+ *
+ * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 only,
+ * as published by the Free Software Foundation.
+ *
+ * This program 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 version 2 for more details (a copy is included
+ * in the LICENSE file that accompanied this code).
+ *
+ * You should have received a copy of the GNU General Public License
+ * version 2 along with this program; If not, see
+ * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
+ *
+ * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
+ * CA 95054 USA or visit www.sun.com if you need additional information or
+ * have any questions.
+ *
+ * GPL HEADER END
+ */
+/*
+ * Copyright (c) 2007, 2010, Oracle and/or its affiliates. All rights reserved.
+ * Use is subject to license terms.
+ */
+/*
+ * This file is part of Lustre, http://www.lustre.org/
+ * Lustre is a trademark of Sun Microsystems, Inc.
*/
#ifndef _LUSTRE_DLM_H__
#define _LUSTRE_DLM_H__
+/** \defgroup ldlm ldlm
+ *
+ * @{
+ */
+
#if defined(__linux__)
#include <linux/lustre_dlm.h>
#elif defined(__APPLE__)
#include <lustre_net.h>
#include <lustre_import.h>
#include <lustre_handles.h>
-#include <lustre_export.h> /* for obd_export, for LDLM_DEBUG */
#include <interval_tree.h> /* for interval_node{}, ldlm_extent */
+#include <lu_ref.h>
struct obd_ops;
struct obd_device;
#define OBD_LDLM_DEVICENAME "ldlm"
-#define LDLM_DEFAULT_LRU_SIZE (100 * num_online_cpus())
+#ifdef HAVE_BGL_SUPPORT
+/* 1.5 times the maximum 128 tasks available in VN mode */
+#define LDLM_DEFAULT_LRU_SIZE 196
+#else
+#define LDLM_DEFAULT_LRU_SIZE (100 * cfs_num_online_cpus())
+#endif
#define LDLM_DEFAULT_MAX_ALIVE (cfs_time_seconds(36000))
#define LDLM_CTIME_AGE_LIMIT (10)
#define LDLM_FL_BLOCK_WAIT 0x000008
#define LDLM_FL_CBPENDING 0x000010 /* this lock is being destroyed */
-#define LDLM_FL_AST_SENT 0x000020 /* blocking or cancel packet was sent */
+#define LDLM_FL_AST_SENT 0x000020 /* blocking or cancel packet was
+ * queued for sending. */
#define LDLM_FL_WAIT_NOREPROC 0x000040 /* not a real flag, not saved in lock */
#define LDLM_FL_CANCEL 0x000080 /* cancellation callback already run */
#define LDLM_FL_REPLAY 0x000100
#define LDLM_FL_INTENT_ONLY 0x000200 /* don't grant lock, just do intent */
-#define LDLM_FL_LOCAL_ONLY 0x000400 /* see ldlm_cli_cancel_unused */
+#define LDLM_FL_LOCAL_ONLY 0x000400
/* don't run the cancel callback under ldlm_cli_cancel_unused */
#define LDLM_FL_FAILED 0x000800
#define LDLM_FL_HAS_INTENT 0x001000 /* lock request has intent */
#define LDLM_FL_CANCELING 0x002000 /* lock cancel has already been sent */
#define LDLM_FL_LOCAL 0x004000 /* local lock (ie, no srv/cli split) */
-#define LDLM_FL_WARN 0x008000 /* see ldlm_cli_cancel_unused */
+/* was LDLM_FL_WARN until 2.0.0 0x008000 */
#define LDLM_FL_DISCARD_DATA 0x010000 /* discard (no writeback) on cancel */
#define LDLM_FL_NO_TIMEOUT 0x020000 /* Blocked by group lock - wait
* list. */
#define LDLM_FL_KMS_IGNORE 0x200000
-/* Don't drop lock covering mmapped file in LRU */
-#define LDLM_FL_NO_LRU 0x400000
-
/* Immediatelly cancel such locks when they block some other locks. Send
- cancel notification to original lock holder, but expect no reply. */
+ * cancel notification to original lock holder, but expect no reply. This is
+ * for clients (like liblustre) that cannot be expected to reliably response
+ * to blocking ast. */
#define LDLM_FL_CANCEL_ON_BLOCK 0x800000
/* Flags flags inherited from parent lock when doing intents. */
#define LDLM_INHERIT_FLAGS (LDLM_FL_CANCEL_ON_BLOCK)
-/* These are flags that are mapped into the flags and ASTs of blocking locks */
-#define LDLM_AST_DISCARD_DATA 0x80000000 /* Add FL_DISCARD to blocking ASTs */
-/* Flags sent in AST lock_flags to be mapped into the receiving lock. */
-#define LDLM_AST_FLAGS (LDLM_FL_DISCARD_DATA)
-
/* completion ast to be executed */
#define LDLM_FL_CP_REQD 0x1000000
* w/o involving separate thread. in order to decrease cs rate */
#define LDLM_FL_ATOMIC_CB 0x4000000
-/* Cancel lock asynchronously. See ldlm_cli_cancel_unused_resource. */
-#define LDLM_FL_ASYNC 0x8000000
+/* was LDLM_FL_ASYNC until 2.0.0 0x8000000 */
/* It may happen that a client initiate 2 operations, e.g. unlink and mkdir,
* such that server send blocking ast for conflict locks to this client for
* the 1st operation, whereas the 2nd operation has canceled this lock and
* is waiting for rpc_lock which is taken by the 1st operation.
* LDLM_FL_BL_AST is to be set by ldlm_callback_handler() to the lock not allow
- * ELC code to cancel it.
+ * ELC code to cancel it.
* LDLM_FL_BL_DONE is to be set by ldlm_cancel_callback() when lock cache is
* droped to let ldlm_callback_handler() return EINVAL to the server. It is
* used when ELC rpc is already prepared and is waiting for rpc_lock, too late
/* measure lock contention and return -EUSERS if locking contention is high */
#define LDLM_FL_DENY_ON_CONTENTION 0x40000000
+/* These are flags that are mapped into the flags and ASTs of blocking locks */
+#define LDLM_AST_DISCARD_DATA 0x80000000 /* Add FL_DISCARD to blocking ASTs */
+
+/* Flags sent in AST lock_flags to be mapped into the receiving lock. */
+#define LDLM_AST_FLAGS (LDLM_FL_DISCARD_DATA)
+
+/*
+ * --------------------------------------------------------------------------
+ * NOTE! Starting from this point, that is, LDLM_FL_* flags with values above
+ * 0x80000000 will not be sent over the wire.
+ * --------------------------------------------------------------------------
+ */
+
+/* Used for marking lock as an target for -EINTR while cp_ast sleep
+ * emulation + race with upcoming bl_ast. */
+#define LDLM_FL_FAIL_LOC 0x100000000ULL
+
+/* Used while processing the unused list to know that we have already
+ * handled this lock and decided to skip it */
+#define LDLM_FL_SKIPPED 0x200000000ULL
+
/* The blocking callback is overloaded to perform two functions. These flags
* indicate which operation should be performed. */
#define LDLM_CB_BLOCKING 1
#define LCK_COMPAT_CR (LCK_COMPAT_CW | LCK_PR | LCK_PW)
#define LCK_COMPAT_NL (LCK_COMPAT_CR | LCK_EX | LCK_GROUP)
#define LCK_COMPAT_GROUP (LCK_GROUP | LCK_NL)
+#define LCK_COMPAT_COS (LCK_COS)
extern ldlm_mode_t lck_compat_array[];
struct ldlm_pool_ops {
int (*po_recalc)(struct ldlm_pool *pl);
- int (*po_shrink)(struct ldlm_pool *pl, int nr,
+ int (*po_shrink)(struct ldlm_pool *pl, int nr,
unsigned int gfp_mask);
int (*po_setup)(struct ldlm_pool *pl, int limit);
};
-/* One second for pools thread check interval. */
+/**
+ * One second for pools thread check interval. Each pool has own period.
+ */
#define LDLM_POOLS_THREAD_PERIOD (1)
-/* 5% margin for modest pools. See ldlm_pool.c for details. */
-#define LDLM_POOLS_MODEST_MARGIN (5)
+/**
+ * ~6% margin for modest pools. See ldlm_pool.c for details.
+ */
+#define LDLM_POOLS_MODEST_MARGIN_SHIFT (4)
+
+/**
+ * Default recalc period for server side pools in sec.
+ */
+#define LDLM_POOL_SRV_DEF_RECALC_PERIOD (1)
-/* A change to SLV in % after which we want to wake up pools thread asap. */
-#define LDLM_POOLS_FAST_SLV_CHANGE (50)
+/**
+ * Default recalc period for client side pools in sec.
+ */
+#define LDLM_POOL_CLI_DEF_RECALC_PERIOD (10)
struct ldlm_pool {
- /**
- * Pool proc directory.
+ /**
+ * Pool proc directory.
*/
cfs_proc_dir_entry_t *pl_proc_dir;
/**
* Pool name, should be long enough to contain compound proc entry name.
*/
char pl_name[100];
- /**
- * Lock for protecting slv/clv updates.
+ /**
+ * Lock for protecting slv/clv updates.
*/
- spinlock_t pl_lock;
+ cfs_spinlock_t pl_lock;
/**
- * Number of allowed locks in in pool, both, client and server side.
+ * Number of allowed locks in in pool, both, client and server side.
*/
- atomic_t pl_limit;
- /**
+ cfs_atomic_t pl_limit;
+ /**
* Number of granted locks in
*/
- atomic_t pl_granted;
- /**
- * Grant rate per T.
+ cfs_atomic_t pl_granted;
+ /**
+ * Grant rate per T.
*/
- atomic_t pl_grant_rate;
- /**
- * Cancel rate per T.
+ cfs_atomic_t pl_grant_rate;
+ /**
+ * Cancel rate per T.
*/
- atomic_t pl_cancel_rate;
- /**
- * Grant speed (GR-CR) per T.
+ cfs_atomic_t pl_cancel_rate;
+ /**
+ * Grant speed (GR-CR) per T.
*/
- atomic_t pl_grant_speed;
- /**
+ cfs_atomic_t pl_grant_speed;
+ /**
* Server lock volume. Protected by pl_lock.
*/
__u64 pl_server_lock_volume;
- /**
+ /**
* Current biggest client lock volume. Protected by pl_lock.
*/
__u64 pl_client_lock_volume;
- /**
+ /**
* Lock volume factor. SLV on client is calculated as following:
* server_slv * lock_volume_factor.
*/
- atomic_t pl_lock_volume_factor;
- /**
- * Time when last slv from server was obtained.
+ cfs_atomic_t pl_lock_volume_factor;
+ /**
+ * Time when last slv from server was obtained.
*/
time_t pl_recalc_time;
/**
- * Recalc and shrink ops.
- */
+ * Recalc period for pool.
+ */
+ time_t pl_recalc_period;
+ /**
+ * Recalc and shrink ops.
+ */
struct ldlm_pool_ops *pl_ops;
/**
- * Planned number of granted locks for next T.
+ * Number of planned locks for next period.
*/
int pl_grant_plan;
- /**
- * Grant plan step for next T.
- */
- int pl_grant_step;
- /**
- * Pool statistics.
+ /**
+ * Pool statistics.
*/
struct lprocfs_stats *pl_stats;
};
void *req_cookie, ldlm_mode_t mode, int flags,
void *data);
+typedef int (*ldlm_cancel_for_recovery)(struct ldlm_lock *lock);
+
struct ldlm_valblock_ops {
int (*lvbo_init)(struct ldlm_resource *res);
- int (*lvbo_update)(struct ldlm_resource *res, struct lustre_msg *m,
- int buf_idx, int increase);
+ int (*lvbo_update)(struct ldlm_resource *res,
+ struct ptlrpc_request *r,
+ int increase);
+ int (*lvbo_free)(struct ldlm_resource *res);
};
typedef enum {
LDLM_NAMESPACE_MODEST = 1 << 1
} ldlm_appetite_t;
-/*
- * Default value for ->ns_shrink_thumb. If lock is not extent one its cost
- * is one page. Here we have 256 pages which is 1M on i386. Thus by default
- * all extent locks which have more than 1M long extent will be kept in lru,
- * others (including ibits locks) will be canceled on memory pressure event.
- */
-#define LDLM_LOCK_SHRINK_THUMB 256
-
-/*
- * Default values for the "max_nolock_size", "contention_time" and
- * "contended_locks" namespace tunables.
+/*
+ * Default values for the "max_nolock_size", "contention_time" and
+ * "contended_locks" namespace tunables.
*/
#define NS_DEFAULT_MAX_NOLOCK_BYTES 0
#define NS_DEFAULT_CONTENTION_SECONDS 2
*/
char *ns_name;
- /**
- * Is this a client-side lock tree?
+ /**
+ * Is this a client-side lock tree?
*/
ldlm_side_t ns_client;
- /**
+ /**
* Namespce connect flags supported by server (may be changed via proc,
* lru resize may be disabled/enabled).
*/
__u64 ns_connect_flags;
- /**
- * Client side orig connect flags supported by server.
+ /**
+ * Client side orig connect flags supported by server.
*/
__u64 ns_orig_connect_flags;
- /**
+ /**
* Hash table for namespace.
*/
- struct list_head *ns_hash;
- spinlock_t ns_hash_lock;
+ cfs_list_t *ns_hash;
+ cfs_spinlock_t ns_hash_lock;
/**
- * Count of resources in the hash.
+ * Count of resources in the hash.
*/
__u32 ns_refcount;
- /**
- * All root resources in namespace.
+ /**
+ * All root resources in namespace.
*/
- struct list_head ns_root_list;
+ cfs_list_t ns_root_list;
- /**
+ /**
* Position in global namespace list.
*/
- struct list_head ns_list_chain;
+ cfs_list_t ns_list_chain;
- /**
- * All root resources in namespace.
+ /**
+ * All root resources in namespace.
*/
- struct list_head ns_unused_list;
+ cfs_list_t ns_unused_list;
int ns_nr_unused;
- spinlock_t ns_unused_lock;
+ cfs_spinlock_t ns_unused_lock;
unsigned int ns_max_unused;
unsigned int ns_max_age;
-
+ unsigned int ns_timeouts;
/**
* Seconds.
*/
unsigned int ns_ctime_age_limit;
-
- /**
- * Lower limit to number of pages in lock to keep it in cache.
- */
- unsigned int ns_shrink_thumb;
/**
* Next debug dump, jiffies.
*/
cfs_time_t ns_next_dump;
- atomic_t ns_locks;
+ cfs_atomic_t ns_locks;
__u64 ns_resources;
ldlm_res_policy ns_policy;
struct ldlm_valblock_ops *ns_lvbo;
struct ldlm_pool ns_pool;
ldlm_appetite_t ns_appetite;
- /**
- * If more than @ns_contented_locks found, the resource considered
- * as contended.
+ /**
+ * If more than \a ns_contended_locks found, the resource is considered
+ * to be contended.
*/
unsigned ns_contended_locks;
- /**
- * The resource remembers contended state during @ns_contention_time,
+ /**
+ * The resource remembers contended state during \a ns_contention_time,
* in seconds.
*/
unsigned ns_contention_time;
- /**
+ /**
* Limit size of nolock requests, in bytes.
*/
unsigned ns_max_nolock_size;
/**
- * Backward link to obd, required for ldlm pool to store new SLV.
+ * Backward link to obd, required for ldlm pool to store new SLV.
*/
struct obd_device *ns_obd;
+
+ struct adaptive_timeout ns_at_estimate;/* estimated lock callback time*/
+
+ /* callback to cancel locks before replaying it during recovery */
+ ldlm_cancel_for_recovery ns_cancel_for_recovery;
};
static inline int ns_is_client(struct ldlm_namespace *ns)
{
LASSERT(ns != NULL);
- LASSERT(!(ns->ns_client & ~(LDLM_NAMESPACE_CLIENT |
+ LASSERT(!(ns->ns_client & ~(LDLM_NAMESPACE_CLIENT |
LDLM_NAMESPACE_SERVER)));
LASSERT(ns->ns_client == LDLM_NAMESPACE_CLIENT ||
ns->ns_client == LDLM_NAMESPACE_SERVER);
static inline int ns_is_server(struct ldlm_namespace *ns)
{
LASSERT(ns != NULL);
- LASSERT(!(ns->ns_client & ~(LDLM_NAMESPACE_CLIENT |
+ LASSERT(!(ns->ns_client & ~(LDLM_NAMESPACE_CLIENT |
LDLM_NAMESPACE_SERVER)));
LASSERT(ns->ns_client == LDLM_NAMESPACE_CLIENT ||
ns->ns_client == LDLM_NAMESPACE_SERVER);
return !!(ns->ns_connect_flags & OBD_CONNECT_LRU_RESIZE);
}
+static inline void ns_register_cancel(struct ldlm_namespace *ns,
+ ldlm_cancel_for_recovery arg)
+{
+ LASSERT(ns != NULL);
+ ns->ns_cancel_for_recovery = arg;
+}
+
/*
*
* Resource hash table
*
*/
-#define RES_HASH_BITS 10
+#define RES_HASH_BITS 12
#define RES_HASH_SIZE (1UL << RES_HASH_BITS)
#define RES_HASH_MASK (RES_HASH_SIZE - 1)
typedef int (*ldlm_completion_callback)(struct ldlm_lock *lock, int flags,
void *data);
typedef int (*ldlm_glimpse_callback)(struct ldlm_lock *lock, void *data);
+typedef unsigned long (*ldlm_weigh_callback)(struct ldlm_lock *lock);
/* Interval node data for each LDLM_EXTENT lock */
struct ldlm_interval {
struct interval_node li_node; /* node for tree mgmt */
- struct list_head li_group; /* the locks which have the same
+ cfs_list_t li_group; /* the locks which have the same
* policy - group of the policy */
};
#define to_ldlm_interval(n) container_of(n, struct ldlm_interval, li_node)
struct interval_node *lit_root; /* actually ldlm_interval */
};
-struct ldlm_lock {
- struct portals_handle l_handle; // must be first in the structure
- atomic_t l_refc;
-
- /* internal spinlock protects l_resource. we should hold this lock
- * first before grabbing res_lock.*/
- spinlock_t l_lock;
-
- /* ldlm_lock_change_resource() can change this */
- struct ldlm_resource *l_resource;
-
- /* protected by ns_hash_lock. FIXME */
- struct list_head l_lru;
-
- /* protected by lr_lock, linkage to resource's lock queues */
- struct list_head l_res_link;
-
- struct ldlm_interval *l_tree_node; /* tree node for ldlm_extent */
+#define LUSTRE_TRACKS_LOCK_EXP_REFS (1)
- /* protected by led_lock */
- struct list_head l_export_chain; // per-export chain of locks
-
- /* protected by lr_lock */
- ldlm_mode_t l_req_mode;
- ldlm_mode_t l_granted_mode;
+/* Cancel flag. */
+typedef enum {
+ LCF_ASYNC = 0x1, /* Cancel locks asynchronously. */
+ LCF_LOCAL = 0x2, /* Cancel locks locally, not notifing server */
+ LCF_BL_AST = 0x4, /* Cancel locks marked as LDLM_FL_BL_AST
+ * in the same RPC */
+} ldlm_cancel_flags_t;
+struct ldlm_lock {
+ /**
+ * Must be first in the structure.
+ */
+ struct portals_handle l_handle;
+ /**
+ * Lock reference count.
+ */
+ cfs_atomic_t l_refc;
+ /**
+ * Internal spinlock protects l_resource. we should hold this lock
+ * first before grabbing res_lock.
+ */
+ cfs_spinlock_t l_lock;
+ /**
+ * ldlm_lock_change_resource() can change this.
+ */
+ struct ldlm_resource *l_resource;
+ /**
+ * Protected by ns_hash_lock. List item for client side lru list.
+ */
+ cfs_list_t l_lru;
+ /**
+ * Protected by lr_lock, linkage to resource's lock queues.
+ */
+ cfs_list_t l_res_link;
+ /**
+ * Tree node for ldlm_extent.
+ */
+ struct ldlm_interval *l_tree_node;
+ /**
+ * Protected by per-bucket exp->exp_lock_hash locks. Per export hash
+ * of locks.
+ */
+ cfs_hlist_node_t l_exp_hash;
+ /**
+ * Protected by lr_lock. Requested mode.
+ */
+ ldlm_mode_t l_req_mode;
+ /**
+ * Granted mode, also protected by lr_lock.
+ */
+ ldlm_mode_t l_granted_mode;
+ /**
+ * Lock enqueue completion handler.
+ */
ldlm_completion_callback l_completion_ast;
+ /**
+ * Lock blocking ast handler.
+ */
ldlm_blocking_callback l_blocking_ast;
+ /**
+ * Lock glimpse handler.
+ */
ldlm_glimpse_callback l_glimpse_ast;
+ ldlm_weigh_callback l_weigh_ast;
- struct obd_export *l_export;
- struct obd_export *l_conn_export;
+ /**
+ * Lock export.
+ */
+ struct obd_export *l_export;
+ /**
+ * Lock connection export.
+ */
+ struct obd_export *l_conn_export;
- struct lustre_handle l_remote_handle;
- ldlm_policy_data_t l_policy_data;
+ /**
+ * Remote lock handle.
+ */
+ struct lustre_handle l_remote_handle;
- /* protected by lr_lock */
- __u32 l_flags;
+ ldlm_policy_data_t l_policy_data;
+
+ /*
+ * Protected by lr_lock. Various counters: readers, writers, etc.
+ */
+ __u64 l_flags;
__u32 l_readers;
__u32 l_writers;
+ /*
+ * Set for locks that were removed from class hash table and will be
+ * destroyed when last reference to them is released. Set by
+ * ldlm_lock_destroy_internal().
+ *
+ * Protected by lock and resource locks.
+ */
__u8 l_destroyed;
- /* If the lock is granted, a process sleeps on this waitq to learn when
+ /**
+ * If the lock is granted, a process sleeps on this waitq to learn when
* it's no longer in use. If the lock is not granted, a process sleeps
- * on this waitq to learn when it becomes granted. */
+ * on this waitq to learn when it becomes granted.
+ */
cfs_waitq_t l_waitq;
- struct timeval l_enqueued_time;
- cfs_time_t l_last_used; /* jiffies */
+ /**
+ * Seconds. it will be updated if there is any activity related to
+ * the lock, e.g. enqueue the lock or send block AST.
+ */
+ cfs_time_t l_last_activity;
+
+ /**
+ * Jiffies. Should be converted to time if needed.
+ */
+ cfs_time_t l_last_used;
+
struct ldlm_extent l_req_extent;
- /* Client-side-only members */
- __u32 l_lvb_len; /* temporary storage for */
- void *l_lvb_data; /* an LVB received during */
- void *l_lvb_swabber; /* an enqueue */
+ /*
+ * Client-side-only members.
+ */
+
+ /**
+ * Temporary storage for an LVB received during an enqueue operation.
+ */
+ __u32 l_lvb_len;
+ void *l_lvb_data;
+
void *l_ast_data;
- spinlock_t l_extents_list_lock;
- struct list_head l_extents_list;
- struct list_head l_cache_locks_list;
+ /*
+ * Server-side-only members.
+ */
- /* Server-side-only members */
+ /** connection cookie for the client originated the operation. */
+ __u64 l_client_cookie;
- /* protected by elt_lock */
- struct list_head l_pending_chain; /* callbacks pending */
- cfs_time_t l_callback_timeout; /* jiffies */
+ /**
+ * Protected by elt_lock. Callbacks pending.
+ */
+ cfs_list_t l_pending_chain;
+
+ cfs_time_t l_callback_timeout;
- __u32 l_pid; /* pid which created this lock */
+ /**
+ * Pid which created this lock.
+ */
+ __u32 l_pid;
+
+ /**
+ * For ldlm_add_ast_work_item().
+ */
+ cfs_list_t l_bl_ast;
+ /**
+ * For ldlm_add_ast_work_item().
+ */
+ cfs_list_t l_cp_ast;
+ /**
+ * For ldlm_add_ast_work_item().
+ */
+ cfs_list_t l_rk_ast;
- /* for ldlm_add_ast_work_item() */
- struct list_head l_bl_ast;
- struct list_head l_cp_ast;
struct ldlm_lock *l_blocking_lock;
int l_bl_ast_run;
- /* protected by lr_lock, linkages to "skip lists" */
- struct list_head l_sl_mode;
- struct list_head l_sl_policy;
+ /**
+ * Protected by lr_lock, linkages to "skip lists".
+ */
+ cfs_list_t l_sl_mode;
+ cfs_list_t l_sl_policy;
+ struct lu_ref l_reference;
+#if LUSTRE_TRACKS_LOCK_EXP_REFS
+ /* Debugging stuff for bug 20498, for tracking export
+ references. */
+ /** number of export references taken */
+ int l_exp_refs_nr;
+ /** link all locks referencing one export */
+ cfs_list_t l_exp_refs_link;
+ /** referenced export object */
+ struct obd_export *l_exp_refs_target;
+#endif
};
struct ldlm_resource {
struct ldlm_namespace *lr_namespace;
/* protected by ns_hash_lock */
- struct list_head lr_hash;
- struct ldlm_resource *lr_parent; /* 0 for a root resource */
- struct list_head lr_children; /* list head for child resources */
- struct list_head lr_childof; /* part of ns_root_list if root res,
+ cfs_list_t lr_hash;
+ cfs_list_t lr_childof; /* part of ns_root_list if root res,
* part of lr_children if child */
- spinlock_t lr_lock;
+ cfs_spinlock_t lr_lock;
/* protected by lr_lock */
- struct list_head lr_granted;
- struct list_head lr_converting;
- struct list_head lr_waiting;
+ cfs_list_t lr_granted;
+ cfs_list_t lr_converting;
+ cfs_list_t lr_waiting;
ldlm_mode_t lr_most_restr;
ldlm_type_t lr_type; /* LDLM_{PLAIN,EXTENT,FLOCK} */
struct ldlm_res_id lr_name;
- atomic_t lr_refcount;
+ cfs_atomic_t lr_refcount;
struct ldlm_interval_tree lr_itree[LCK_MODE_NUM]; /* interval trees*/
/* Server-side-only lock value block elements */
- struct semaphore lr_lvb_sem;
+ cfs_semaphore_t lr_lvb_sem;
__u32 lr_lvb_len;
void *lr_lvb_data;
/* when the resource was considered as contended */
cfs_time_t lr_contention_time;
+ /**
+ * List of references to this resource. For debugging.
+ */
+ struct lu_ref lr_reference;
+
+ struct inode *lr_lvb_inode;
};
+static inline struct ldlm_namespace *
+ldlm_res_to_ns(struct ldlm_resource *res)
+{
+ return res->lr_namespace;
+}
+
+static inline struct ldlm_namespace *
+ldlm_lock_to_ns(struct ldlm_lock *lock)
+{
+ return ldlm_res_to_ns(lock->l_resource);
+}
+
+static inline char *
+ldlm_lock_to_ns_name(struct ldlm_lock *lock)
+{
+ return ldlm_lock_to_ns(lock)->ns_name;
+}
+
+static inline struct adaptive_timeout *
+ldlm_lock_to_ns_at(struct ldlm_lock *lock)
+{
+ return &ldlm_lock_to_ns(lock)->ns_at_estimate;
+}
+
struct ldlm_ast_work {
- struct ldlm_lock *w_lock;
- int w_blocking;
- struct ldlm_lock_desc w_desc;
- struct list_head w_list;
- int w_flags;
- void *w_data;
- int w_datalen;
+ struct ldlm_lock *w_lock;
+ int w_blocking;
+ struct ldlm_lock_desc w_desc;
+ cfs_list_t w_list;
+ int w_flags;
+ void *w_data;
+ int w_datalen;
};
/* ldlm_enqueue parameters common */
void *ei_cb_bl; /* blocking lock callback */
void *ei_cb_cp; /* lock completion callback */
void *ei_cb_gl; /* lock glimpse callback */
+ void *ei_cb_wg; /* lock weigh callback */
void *ei_cbdata; /* Data to be passed into callbacks. */
+ short ei_async:1; /* async request */
};
extern struct obd_ops ldlm_obd_ops;
extern char *ldlm_it2str(int it);
#ifdef LIBCFS_DEBUG
#define ldlm_lock_debug(cdls, level, lock, file, func, line, fmt, a...) do { \
- CHECK_STACK(); \
+ CFS_CHECK_STACK(); \
\
if (((level) & D_CANTMASK) != 0 || \
((libcfs_debug & (level)) != 0 && \
...)
__attribute__ ((format (printf, 4, 5)));
-#define LDLM_ERROR(lock, fmt, a...) do { \
+#define LDLM_DEBUG_LIMIT(mask, lock, fmt, a...) do { \
static cfs_debug_limit_state_t _ldlm_cdls; \
- ldlm_lock_debug(&_ldlm_cdls, D_ERROR, lock, \
+ ldlm_lock_debug(&_ldlm_cdls, mask, lock, \
__FILE__, __FUNCTION__, __LINE__, \
"### " fmt , ##a); \
} while (0)
+#define LDLM_ERROR(lock, fmt, a...) LDLM_DEBUG_LIMIT(D_ERROR, lock, fmt, ## a)
+#define LDLM_WARN(lock, fmt, a...) LDLM_DEBUG_LIMIT(D_WARNING, lock, fmt, ## a)
+
#define LDLM_DEBUG(lock, fmt, a...) do { \
ldlm_lock_debug(NULL, D_DLMTRACE, lock, \
__FILE__, __FUNCTION__, __LINE__, \
"### " fmt , ##a); \
} while (0)
-#else
-#define LDLM_DEBUG(lock, fmt, a...) ((void)0)
-#define LDLM_ERROR(lock, fmt, a...) ((void)0)
+#else /* !LIBCFS_DEBUG */
+# define LDLM_DEBUG(lock, fmt, a...) ((void)0)
+# define LDLM_ERROR(lock, fmt, a...) ((void)0)
+# define ldlm_lock_debuf(cdls, level, lock, file, func, line, fmt, a...) \
+ ((void)0)
#endif
#define LDLM_DEBUG_NOLOCK(format, a...) \
typedef int (*ldlm_processing_policy)(struct ldlm_lock *lock, int *flags,
int first_enq, ldlm_error_t *err,
- struct list_head *work_list);
+ cfs_list_t *work_list);
/*
* Iterators.
ldlm_res_iterator_t iter, void *closure);
int ldlm_replay_locks(struct obd_import *imp);
-void ldlm_resource_iterate(struct ldlm_namespace *, const struct ldlm_res_id *,
+int ldlm_resource_iterate(struct ldlm_namespace *, const struct ldlm_res_id *,
ldlm_iterator_t iter, void *data);
/* ldlm_flock.c */
int ldlm_request_cancel(struct ptlrpc_request *req,
const struct ldlm_request *dlm_req, int first);
int ldlm_del_waiting_lock(struct ldlm_lock *lock);
-int ldlm_refresh_waiting_lock(struct ldlm_lock *lock);
+int ldlm_refresh_waiting_lock(struct ldlm_lock *lock, int timeout);
void ldlm_revoke_export_locks(struct obd_export *exp);
int ldlm_get_ref(void);
void ldlm_put_ref(void);
+int ldlm_init_export(struct obd_export *exp);
+void ldlm_destroy_export(struct obd_export *exp);
/* ldlm_lock.c */
ldlm_processing_policy ldlm_get_processing_policy(struct ldlm_resource *res);
struct lustre_handle *lockh);
struct ldlm_lock *__ldlm_handle2lock(const struct lustre_handle *, int flags);
void ldlm_cancel_callback(struct ldlm_lock *);
-int ldlm_lock_set_data(struct lustre_handle *, void *data);
int ldlm_lock_remove_from_lru(struct ldlm_lock *);
-struct ldlm_lock *ldlm_handle2lock_ns(struct ldlm_namespace *,
- const struct lustre_handle *);
static inline struct ldlm_lock *ldlm_handle2lock(const struct lustre_handle *h)
{
return __ldlm_handle2lock(h, 0);
}
+#define LDLM_LOCK_REF_DEL(lock) \
+ lu_ref_del(&lock->l_reference, "handle", cfs_current())
+
+static inline struct ldlm_lock *
+ldlm_handle2lock_long(const struct lustre_handle *h, int flags)
+{
+ struct ldlm_lock *lock;
+
+ lock = __ldlm_handle2lock(h, flags);
+ if (lock != NULL)
+ LDLM_LOCK_REF_DEL(lock);
+ return lock;
+}
+
static inline int ldlm_res_lvbo_update(struct ldlm_resource *res,
- struct lustre_msg *m, int buf_idx,
- int increase)
+ struct ptlrpc_request *r, int increase)
{
- if (res->lr_namespace->ns_lvbo &&
- res->lr_namespace->ns_lvbo->lvbo_update) {
- return res->lr_namespace->ns_lvbo->lvbo_update(res, m, buf_idx,
- increase);
+ if (ldlm_res_to_ns(res)->ns_lvbo &&
+ ldlm_res_to_ns(res)->ns_lvbo->lvbo_update) {
+ return ldlm_res_to_ns(res)->ns_lvbo->lvbo_update(res, r,
+ increase);
}
return 0;
}
+int ldlm_error2errno(ldlm_error_t error);
+ldlm_error_t ldlm_errno2error(int err_no); /* don't call it `errno': this
+ * confuses user-space. */
+#if LUSTRE_TRACKS_LOCK_EXP_REFS
+void ldlm_dump_export_locks(struct obd_export *exp);
+#endif
+
+/**
+ * Release a temporary lock reference obtained by ldlm_handle2lock() or
+ * __ldlm_handle2lock().
+ */
#define LDLM_LOCK_PUT(lock) \
do { \
+ LDLM_LOCK_REF_DEL(lock); \
+ /*LDLM_DEBUG((lock), "put");*/ \
+ ldlm_lock_put(lock); \
+} while (0)
+
+/**
+ * Release a lock reference obtained by some other means (see
+ * LDLM_LOCK_PUT()).
+ */
+#define LDLM_LOCK_RELEASE(lock) \
+do { \
/*LDLM_DEBUG((lock), "put");*/ \
ldlm_lock_put(lock); \
} while (0)
lock; \
})
-#define ldlm_lock_list_put(head, member, count) \
-({ \
- struct ldlm_lock *_lock, *_next; \
- int c = count; \
- list_for_each_entry_safe(_lock, _next, head, member) { \
- if (c-- == 0) \
- break; \
- list_del_init(&_lock->member); \
- LDLM_LOCK_PUT(_lock); \
- } \
- LASSERT(c <= 0); \
+#define ldlm_lock_list_put(head, member, count) \
+({ \
+ struct ldlm_lock *_lock, *_next; \
+ int c = count; \
+ cfs_list_for_each_entry_safe(_lock, _next, head, member) { \
+ if (c-- == 0) \
+ break; \
+ cfs_list_del_init(&_lock->member); \
+ LDLM_LOCK_RELEASE(_lock); \
+ } \
+ LASSERT(c <= 0); \
})
struct ldlm_lock *ldlm_lock_get(struct ldlm_lock *lock);
void ldlm_lock_destroy(struct ldlm_lock *lock);
void ldlm_lock2desc(struct ldlm_lock *lock, struct ldlm_lock_desc *desc);
void ldlm_lock_addref(struct lustre_handle *lockh, __u32 mode);
+int ldlm_lock_addref_try(struct lustre_handle *lockh, __u32 mode);
void ldlm_lock_decref(struct lustre_handle *lockh, __u32 mode);
void ldlm_lock_decref_and_cancel(struct lustre_handle *lockh, __u32 mode);
void ldlm_lock_allow_match(struct ldlm_lock *lock);
-int ldlm_lock_fast_match(struct ldlm_lock *, int, obd_off, obd_off, void **);
-void ldlm_lock_fast_release(void *, int);
+void ldlm_lock_allow_match_locked(struct ldlm_lock *lock);
ldlm_mode_t ldlm_lock_match(struct ldlm_namespace *ns, int flags,
const struct ldlm_res_id *, ldlm_type_t type,
ldlm_policy_data_t *, ldlm_mode_t mode,
- struct lustre_handle *);
+ struct lustre_handle *, int unref);
struct ldlm_resource *ldlm_lock_convert(struct ldlm_lock *lock, int new_mode,
__u32 *flags);
+void ldlm_lock_downgrade(struct ldlm_lock *lock, int new_mode);
void ldlm_lock_cancel(struct ldlm_lock *lock);
-void ldlm_cancel_locks_for_export(struct obd_export *export);
void ldlm_reprocess_all(struct ldlm_resource *res);
void ldlm_reprocess_all_ns(struct ldlm_namespace *ns);
void ldlm_lock_dump(int level, struct ldlm_lock *lock, int pos);
/* resource.c */
struct ldlm_namespace *
-ldlm_namespace_new(struct obd_device *obd, char *name,
+ldlm_namespace_new(struct obd_device *obd, char *name,
ldlm_side_t client, ldlm_appetite_t apt);
int ldlm_namespace_cleanup(struct ldlm_namespace *ns, int flags);
-void ldlm_namespace_free(struct ldlm_namespace *ns,
+void ldlm_namespace_free(struct ldlm_namespace *ns,
struct obd_import *imp, int force);
void ldlm_namespace_register(struct ldlm_namespace *ns, ldlm_side_t client);
void ldlm_namespace_unregister(struct ldlm_namespace *ns, ldlm_side_t client);
ldlm_type_t type, int create);
struct ldlm_resource *ldlm_resource_getref(struct ldlm_resource *res);
int ldlm_resource_putref(struct ldlm_resource *res);
-void ldlm_resource_add_lock(struct ldlm_resource *res, struct list_head *head,
+void ldlm_resource_add_lock(struct ldlm_resource *res,
+ cfs_list_t *head,
struct ldlm_lock *lock);
void ldlm_resource_unlink_lock(struct ldlm_lock *lock);
void ldlm_res2desc(struct ldlm_resource *res, struct ldlm_resource_desc *desc);
int ldlm_lock_change_resource(struct ldlm_namespace *, struct ldlm_lock *,
const struct ldlm_res_id *);
+#define LDLM_RESOURCE_ADDREF(res) do { \
+ lu_ref_add_atomic(&(res)->lr_reference, __FUNCTION__, cfs_current()); \
+} while (0)
+
+#define LDLM_RESOURCE_DELREF(res) do { \
+ lu_ref_del(&(res)->lr_reference, __FUNCTION__, cfs_current()); \
+} while (0)
+
struct ldlm_callback_suite {
ldlm_completion_callback lcs_completion;
ldlm_blocking_callback lcs_blocking;
ldlm_glimpse_callback lcs_glimpse;
+ ldlm_weigh_callback lcs_weigh;
};
/* ldlm_request.c */
int ldlm_expired_completion_wait(void *data);
+int ldlm_blocking_ast_nocheck(struct ldlm_lock *lock);
int ldlm_blocking_ast(struct ldlm_lock *lock, struct ldlm_lock_desc *desc,
void *data, int flag);
int ldlm_glimpse_ast(struct ldlm_lock *lock, void *reqp);
+int ldlm_completion_ast_async(struct ldlm_lock *lock, int flags, void *data);
int ldlm_completion_ast(struct ldlm_lock *lock, int flags, void *data);
int ldlm_cli_enqueue(struct obd_export *exp, struct ptlrpc_request **reqp,
struct ldlm_enqueue_info *einfo,
const struct ldlm_res_id *res_id,
- ldlm_policy_data_t *policy, int *flags,
- void *lvb, __u32 lvb_len, void *lvb_swabber,
- struct lustre_handle *lockh, int async);
+ ldlm_policy_data_t const *policy, int *flags,
+ void *lvb, __u32 lvb_len, struct lustre_handle *lockh,
+ int async);
int ldlm_prep_enqueue_req(struct obd_export *exp,
struct ptlrpc_request *req,
- struct list_head *cancels,
+ cfs_list_t *cancels,
int count);
int ldlm_prep_elc_req(struct obd_export *exp,
struct ptlrpc_request *req,
int version, int opc, int canceloff,
- struct list_head *cancels, int count);
+ cfs_list_t *cancels, int count);
int ldlm_handle_enqueue0(struct ldlm_namespace *ns, struct ptlrpc_request *req,
const struct ldlm_request *dlm_req,
const struct ldlm_callback_suite *cbs);
int ldlm_cli_enqueue_fini(struct obd_export *exp, struct ptlrpc_request *req,
ldlm_type_t type, __u8 with_policy, ldlm_mode_t mode,
int *flags, void *lvb, __u32 lvb_len,
- void *lvb_swabber, struct lustre_handle *lockh,
- int rc);
+ struct lustre_handle *lockh, int rc);
int ldlm_cli_enqueue_local(struct ldlm_namespace *ns,
const struct ldlm_res_id *res_id,
ldlm_type_t type, ldlm_policy_data_t *policy,
ldlm_blocking_callback blocking,
ldlm_completion_callback completion,
ldlm_glimpse_callback glimpse,
- void *data, __u32 lvb_len, void *lvb_swabber,
+ void *data, __u32 lvb_len,
+ const __u64 *client_cookie,
struct lustre_handle *lockh);
int ldlm_server_ast(struct lustre_handle *lockh, struct ldlm_lock_desc *new,
void *data, __u32 data_len);
const struct ldlm_request *dlm_req);
int ldlm_cli_cancel(struct lustre_handle *lockh);
int ldlm_cli_cancel_unused(struct ldlm_namespace *, const struct ldlm_res_id *,
- int flags, void *opaque);
+ ldlm_cancel_flags_t flags, void *opaque);
int ldlm_cli_cancel_unused_resource(struct ldlm_namespace *ns,
const struct ldlm_res_id *res_id,
ldlm_policy_data_t *policy,
- ldlm_mode_t mode, int flags, void *opaque);
-int ldlm_cli_cancel_req(struct obd_export *exp, struct list_head *head,
- int count, int flags);
-int ldlm_cli_join_lru(struct ldlm_namespace *,
- const struct ldlm_res_id *, int join);
+ ldlm_mode_t mode,
+ ldlm_cancel_flags_t flags,
+ void *opaque);
+int ldlm_cli_cancel_req(struct obd_export *exp, cfs_list_t *head,
+ int count, ldlm_cancel_flags_t flags);
int ldlm_cancel_resource_local(struct ldlm_resource *res,
- struct list_head *cancels,
+ cfs_list_t *cancels,
ldlm_policy_data_t *policy,
ldlm_mode_t mode, int lock_flags,
- int cancel_flags, void *opaque);
-int ldlm_cli_cancel_list(struct list_head *head, int count,
- struct ptlrpc_request *req, int flags);
-
+ ldlm_cancel_flags_t cancel_flags, void *opaque);
+int ldlm_cli_cancel_list_local(cfs_list_t *cancels, int count,
+ ldlm_cancel_flags_t flags);
+int ldlm_cli_cancel_list(cfs_list_t *head, int count,
+ struct ptlrpc_request *req, ldlm_cancel_flags_t flags);
+
/* mds/handler.c */
/* This has to be here because recursive inclusion sucks. */
int intent_disposition(struct ldlm_reply *rep, int flag);
#define IOC_LDLM_REGRESS_STOP _IOWR('f', 43, long)
#define IOC_LDLM_MAX_NR 43
+/**
+ * "Modes" of acquiring lock_res, necessary to tell lockdep that taking more
+ * than one lock_res is dead-lock safe.
+ */
+enum lock_res_type {
+ LRT_NORMAL,
+ LRT_NEW
+};
+
static inline void lock_res(struct ldlm_resource *res)
{
- spin_lock(&res->lr_lock);
+ cfs_spin_lock(&res->lr_lock);
}
+static inline void lock_res_nested(struct ldlm_resource *res,
+ enum lock_res_type mode)
+{
+ cfs_spin_lock_nested(&res->lr_lock, mode);
+}
+
+
static inline void unlock_res(struct ldlm_resource *res)
{
- spin_unlock(&res->lr_lock);
+ cfs_spin_unlock(&res->lr_lock);
}
static inline void check_res_locked(struct ldlm_resource *res)
void ldlm_pools_recalc(ldlm_side_t client);
int ldlm_pools_init(void);
void ldlm_pools_fini(void);
-void ldlm_pools_wakeup(void);
-int ldlm_pool_init(struct ldlm_pool *pl, struct ldlm_namespace *ns,
+int ldlm_pool_init(struct ldlm_pool *pl, struct ldlm_namespace *ns,
int idx, ldlm_side_t client);
-int ldlm_pool_shrink(struct ldlm_pool *pl, int nr,
+int ldlm_pool_shrink(struct ldlm_pool *pl, int nr,
unsigned int gfp_mask);
void ldlm_pool_fini(struct ldlm_pool *pl);
int ldlm_pool_setup(struct ldlm_pool *pl, int limit);
void ldlm_pool_set_limit(struct ldlm_pool *pl, __u32 limit);
void ldlm_pool_add(struct ldlm_pool *pl, struct ldlm_lock *lock);
void ldlm_pool_del(struct ldlm_pool *pl, struct ldlm_lock *lock);
-#endif
+/** @} ldlm */
+
+#endif
git://git.whamcloud.com / fs / lustre-release.git / blobdiff