path: root/drivers/staging/lustre/lustre/obdclass/cl_object.c

// SPDX-License-Identifier: GPL-2.0
/*
 * 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.gnu.org/licenses/gpl-2.0.html
 *
 * GPL HEADER END
 */
/*
 * Copyright (c) 2008, 2010, Oracle and/or its affiliates. All rights reserved.
 * Use is subject to license terms.
 *
 * Copyright (c) 2011, 2015, Intel Corporation.
 */
/*
 * This file is part of Lustre, http://www.lustre.org/
 * Lustre is a trademark of Sun Microsystems, Inc.
 *
 * Client Lustre Object.
 *
 *   Author: Nikita Danilov <nikita.danilov@sun.com>
 *   Author: Jinshan Xiong <jinshan.xiong@intel.com>
 */

/*
 * Locking.
 *
 *  i_mutex
 *      PG_locked
 *	  ->coh_attr_guard
 *	  ->ls_guard
 */

#define DEBUG_SUBSYSTEM S_CLASS

#include <linux/libcfs/libcfs.h>
/* class_put_type() */
#include <obd_class.h>
#include <obd_support.h>
#include <lustre_fid.h>
#include <linux/list.h>
#include <linux/libcfs/libcfs_hash.h>	/* for cfs_hash stuff */
#include <cl_object.h>
#include <lu_object.h>
#include "cl_internal.h"

static struct kmem_cache *cl_env_kmem;

/** Lock class of cl_object_header::coh_attr_guard */
static struct lock_class_key cl_attr_guard_class;

/**
 * Initialize cl_object_header.
 */
int cl_object_header_init(struct cl_object_header *h)
{
	int result;

	result = lu_object_header_init(&h->coh_lu);
	if (result == 0) {
		spin_lock_init(&h->coh_attr_guard);
		lockdep_set_class(&h->coh_attr_guard, &cl_attr_guard_class);
		h->coh_page_bufsize = 0;
	}
	return result;
}
EXPORT_SYMBOL(cl_object_header_init);

/**
 * Returns a cl_object with a given \a fid.
 *
 * Returns either cached or newly created object. Additional reference on the
 * returned object is acquired.
 *
 * \see lu_object_find(), cl_page_find(), cl_lock_find()
 */
struct cl_object *cl_object_find(const struct lu_env *env,
				 struct cl_device *cd, const struct lu_fid *fid,
				 const struct cl_object_conf *c)
{
	might_sleep();
	return lu2cl(lu_object_find_slice(env, cl2lu_dev(cd), fid, &c->coc_lu));
}
EXPORT_SYMBOL(cl_object_find);

/**
 * Releases a reference on \a o.
 *
 * When last reference is released object is returned to the cache, unless
 * lu_object_header_flags::LU_OBJECT_HEARD_BANSHEE bit is set in its header.
 *
 * \see cl_page_put(), cl_lock_put().
 */
void cl_object_put(const struct lu_env *env, struct cl_object *o)
{
	lu_object_put(env, &o->co_lu);
}
EXPORT_SYMBOL(cl_object_put);

/**
 * Acquire an additional reference to the object \a o.
 *
 * This can only be used to acquire _additional_ reference, i.e., caller
 * already has to possess at least one reference to \a o before calling this.
 *
 * \see cl_page_get(), cl_lock_get().
 */
void cl_object_get(struct cl_object *o)
{
	lu_object_get(&o->co_lu);
}
EXPORT_SYMBOL(cl_object_get);

/**
 * Returns the top-object for a given \a o.
 *
 * \see cl_io_top()
 */
struct cl_object *cl_object_top(struct cl_object *o)
{
	struct cl_object_header *hdr = cl_object_header(o);
	struct cl_object *top;

	while (hdr->coh_parent)
		hdr = hdr->coh_parent;

	top = lu2cl(lu_object_top(&hdr->coh_lu));
	CDEBUG(D_TRACE, "%p -> %p\n", o, top);
	return top;
}
EXPORT_SYMBOL(cl_object_top);

/**
 * Returns pointer to the lock protecting data-attributes for the given object
 * \a o.
 *
 * Data-attributes are protected by the cl_object_header::coh_attr_guard
 * spin-lock in the top-object.
 *
 * \see cl_attr, cl_object_attr_lock(), cl_object_operations::coo_attr_get().
 */
static spinlock_t *cl_object_attr_guard(struct cl_object *o)
{
	return &cl_object_header(cl_object_top(o))->coh_attr_guard;
}

/**
 * Locks data-attributes.
 *
 * Prevents data-attributes from changing, until lock is released by
 * cl_object_attr_unlock(). This has to be called before calls to
 * cl_object_attr_get(), cl_object_attr_update().
 */
void cl_object_attr_lock(struct cl_object *o)
	__acquires(cl_object_attr_guard(o))
{
	spin_lock(cl_object_attr_guard(o));
}
EXPORT_SYMBOL(cl_object_attr_lock);

/**
 * Releases data-attributes lock, acquired by cl_object_attr_lock().
 */
void cl_object_attr_unlock(struct cl_object *o)
	__releases(cl_object_attr_guard(o))
{
	spin_unlock(cl_object_attr_guard(o));
}
EXPORT_SYMBOL(cl_object_attr_unlock);

/**
 * Returns data-attributes of an object \a obj.
 *
 * Every layer is asked (by calling cl_object_operations::coo_attr_get())
 * top-to-bottom to fill in parts of \a attr that this layer is responsible
 * for.
 */
int cl_object_attr_get(const struct lu_env *env, struct cl_object *obj,
		       struct cl_attr *attr)
{
	struct lu_object_header *top;
	int result;

	assert_spin_locked(cl_object_attr_guard(obj));

	top = obj->co_lu.lo_header;
	result = 0;
	list_for_each_entry(obj, &top->loh_layers, co_lu.lo_linkage) {
		if (obj->co_ops->coo_attr_get) {
			result = obj->co_ops->coo_attr_get(env, obj, attr);
			if (result != 0) {
				if (result > 0)
					result = 0;
				break;
			}
		}
	}
	return result;
}
EXPORT_SYMBOL(cl_object_attr_get);

/**
 * Updates data-attributes of an object \a obj.
 *
 * Only attributes, mentioned in a validness bit-mask \a v are
 * updated. Calls cl_object_operations::coo_attr_update() on every layer,
 * bottom to top.
 */
int cl_object_attr_update(const struct lu_env *env, struct cl_object *obj,
			  const struct cl_attr *attr, unsigned int v)
{
	struct lu_object_header *top;
	int result;

	assert_spin_locked(cl_object_attr_guard(obj));

	top = obj->co_lu.lo_header;
	result = 0;
	list_for_each_entry_reverse(obj, &top->loh_layers, co_lu.lo_linkage) {
		if (obj->co_ops->coo_attr_update) {
			result = obj->co_ops->coo_attr_update(env, obj, attr,
							      v);
			if (result != 0) {
				if (result > 0)
					result = 0;
				break;
			}
		}
	}
	return result;
}
EXPORT_SYMBOL(cl_object_attr_update);

/**
 * Notifies layers (bottom-to-top) that glimpse AST was received.
 *
 * Layers have to fill \a lvb fields with information that will be shipped
 * back to glimpse issuer.
 *
 * \see cl_lock_operations::clo_glimpse()
 */
int cl_object_glimpse(const struct lu_env *env, struct cl_object *obj,
		      struct ost_lvb *lvb)
{
	struct lu_object_header *top;
	int result;

	top = obj->co_lu.lo_header;
	result = 0;
	list_for_each_entry_reverse(obj, &top->loh_layers, co_lu.lo_linkage) {
		if (obj->co_ops->coo_glimpse) {
			result = obj->co_ops->coo_glimpse(env, obj, lvb);
			if (result != 0)
				break;
		}
	}
	LU_OBJECT_HEADER(D_DLMTRACE, env, lu_object_top(top),
			 "size: %llu mtime: %llu atime: %llu ctime: %llu blocks: %llu\n",
			 lvb->lvb_size, lvb->lvb_mtime, lvb->lvb_atime,
			 lvb->lvb_ctime, lvb->lvb_blocks);
	return result;
}
EXPORT_SYMBOL(cl_object_glimpse);

/**
 * Updates a configuration of an object \a obj.
 */
int cl_conf_set(const struct lu_env *env, struct cl_object *obj,
		const struct cl_object_conf *conf)
{
	struct lu_object_header *top;
	int result;

	top = obj->co_lu.lo_header;
	result = 0;
	list_for_each_entry(obj, &top->loh_layers, co_lu.lo_linkage) {
		if (obj->co_ops->coo_conf_set) {
			result = obj->co_ops->coo_conf_set(env, obj, conf);
			if (result != 0)
				break;
		}
	}
	return result;
}
EXPORT_SYMBOL(cl_conf_set);

/**
 * Prunes caches of pages and locks for this object.
 */
int cl_object_prune(const struct lu_env *env, struct cl_object *obj)
{
	struct lu_object_header *top;
	struct cl_object *o;
	int result;

	top = obj->co_lu.lo_header;
	result = 0;
	list_for_each_entry(o, &top->loh_layers, co_lu.lo_linkage) {
		if (o->co_ops->coo_prune) {
			result = o->co_ops->coo_prune(env, o);
			if (result != 0)
				break;
		}
	}

	return result;
}
EXPORT_SYMBOL(cl_object_prune);

/**
 * Get stripe information of this object.
 */
int cl_object_getstripe(const struct lu_env *env, struct cl_object *obj,
			struct lov_user_md __user *uarg)
{
	struct lu_object_header *top;
	int result = 0;

	top = obj->co_lu.lo_header;
	list_for_each_entry(obj, &top->loh_layers, co_lu.lo_linkage) {
		if (obj->co_ops->coo_getstripe) {
			result = obj->co_ops->coo_getstripe(env, obj, uarg);
			if (result)
				break;
		}
	}
	return result;
}
EXPORT_SYMBOL(cl_object_getstripe);

/**
 * Get fiemap extents from file object.
 *
 * \param env [in]	lustre environment
 * \param obj [in]	file object
 * \param key [in]	fiemap request argument
 * \param fiemap [out]	fiemap extents mapping retrived
 * \param buflen [in]	max buffer length of @fiemap
 *
 * \retval 0	success
 * \retval < 0	error
 */
int cl_object_fiemap(const struct lu_env *env, struct cl_object *obj,
		     struct ll_fiemap_info_key *key,
		     struct fiemap *fiemap, size_t *buflen)
{
	struct lu_object_header *top;
	int result = 0;

	top = obj->co_lu.lo_header;
	list_for_each_entry(obj, &top->loh_layers, co_lu.lo_linkage) {
		if (obj->co_ops->coo_fiemap) {
			result = obj->co_ops->coo_fiemap(env, obj, key, fiemap,
							 buflen);
			if (result)
				break;
		}
	}
	return result;
}
EXPORT_SYMBOL(cl_object_fiemap);

int cl_object_layout_get(const struct lu_env *env, struct cl_object *obj,
			 struct cl_layout *cl)
{
	struct lu_object_header *top = obj->co_lu.lo_header;

	list_for_each_entry(obj, &top->loh_layers, co_lu.lo_linkage) {
		if (obj->co_ops->coo_layout_get)
			return obj->co_ops->coo_layout_get(env, obj, cl);
	}

	return -EOPNOTSUPP;
}
EXPORT_SYMBOL(cl_object_layout_get);

loff_t cl_object_maxbytes(struct cl_object *obj)
{
	struct lu_object_header *top = obj->co_lu.lo_header;
	loff_t maxbytes = LLONG_MAX;

	list_for_each_entry(obj, &top->loh_layers, co_lu.lo_linkage) {
		if (obj->co_ops->coo_maxbytes)
			maxbytes = min_t(loff_t, obj->co_ops->coo_maxbytes(obj),
					 maxbytes);
	}

	return maxbytes;
}
EXPORT_SYMBOL(cl_object_maxbytes);

/**
 * Helper function removing all object locks, and marking object for
 * deletion. All object pages must have been deleted at this point.
 *
 * This is called by cl_inode_fini() and lov_object_delete() to destroy top-
 * and sub- objects respectively.
 */
void cl_object_kill(const struct lu_env *env, struct cl_object *obj)
{
	struct cl_object_header *hdr = cl_object_header(obj);

	set_bit(LU_OBJECT_HEARD_BANSHEE, &hdr->coh_lu.loh_flags);
}
EXPORT_SYMBOL(cl_object_kill);

void cache_stats_init(struct cache_stats *cs, const char *name)
{
	int i;

	cs->cs_name = name;
	for (i = 0; i < CS_NR; i++)
		atomic_set(&cs->cs_stats[i], 0);
}

static int cache_stats_print(const struct cache_stats *cs,
			     struct seq_file *m, int h)
{
	int i;
	/*
	 *   lookup    hit    total  cached create
	 * env: ...... ...... ...... ...... ......
	 */
	if (h) {
		const char *names[CS_NR] = CS_NAMES;

		seq_printf(m, "%6s", " ");
		for (i = 0; i < CS_NR; i++)
			seq_printf(m, "%8s", names[i]);
		seq_printf(m, "\n");
	}

	seq_printf(m, "%5.5s:", cs->cs_name);
	for (i = 0; i < CS_NR; i++)
		seq_printf(m, "%8u", atomic_read(&cs->cs_stats[i]));
	return 0;
}

static void cl_env_percpu_refill(void);

/**
 * Initialize client site.
 *
 * Perform common initialization (lu_site_init()), and initialize statistical
 * counters. Also perform global initializations on the first call.
 */
int cl_site_init(struct cl_site *s, struct cl_device *d)
{
	size_t i;
	int result;

	result = lu_site_init(&s->cs_lu, &d->cd_lu_dev);
	if (result == 0) {
		cache_stats_init(&s->cs_pages, "pages");
		for (i = 0; i < ARRAY_SIZE(s->cs_pages_state); ++i)
			atomic_set(&s->cs_pages_state[0], 0);
		cl_env_percpu_refill();
	}
	return result;
}
EXPORT_SYMBOL(cl_site_init);

/**
 * Finalize client site. Dual to cl_site_init().
 */
void cl_site_fini(struct cl_site *s)
{
	lu_site_fini(&s->cs_lu);
}
EXPORT_SYMBOL(cl_site_fini);

static struct cache_stats cl_env_stats = {
	.cs_name    = "envs",
	.cs_stats = { ATOMIC_INIT(0), }
};

/**
 * Outputs client site statistical counters into a buffer. Suitable for
 * ll_rd_*()-style functions.
 */
int cl_site_stats_print(const struct cl_site *site, struct seq_file *m)
{
	size_t i;
	static const char *pstate[] = {
		[CPS_CACHED]  = "c",
		[CPS_OWNED]   = "o",
		[CPS_PAGEOUT] = "w",
		[CPS_PAGEIN]  = "r",
		[CPS_FREEING] = "f"
	};
/*
       lookup    hit  total   busy create
pages: ...... ...... ...... ...... ...... [...... ...... ...... ......]
locks: ...... ...... ...... ...... ...... [...... ...... ...... ...... ......]
  env: ...... ...... ...... ...... ......
 */
	lu_site_stats_print(&site->cs_lu, m);
	cache_stats_print(&site->cs_pages, m, 1);
	seq_printf(m, " [");
	for (i = 0; i < ARRAY_SIZE(site->cs_pages_state); ++i)
		seq_printf(m, "%s: %u ", pstate[i],
			   atomic_read(&site->cs_pages_state[i]));
	seq_printf(m, "]\n");
	cache_stats_print(&cl_env_stats, m, 0);
	seq_printf(m, "\n");
	return 0;
}
EXPORT_SYMBOL(cl_site_stats_print);

/*****************************************************************************
 *
 * lu_env handling on client.
 *
 */

/**
 * The most efficient way is to store cl_env pointer in task specific
 * structures. On Linux, it wont' be easy to use task_struct->journal_info
 * because Lustre code may call into other fs which has certain assumptions
 * about journal_info. Currently following fields in task_struct are identified
 * can be used for this purpose:
 *  - tux_info: only on RedHat kernel.
 *  - ...
 * \note As long as we use task_struct to store cl_env, we assume that once
 * called into Lustre, we'll never call into the other part of the kernel
 * which will use those fields in task_struct without explicitly exiting
 * Lustre.
 *
 * If there's no space in task_struct is available, hash will be used.
 * bz20044, bz22683.
 */

static unsigned int cl_envs_cached_max = 32; /* XXX: prototype: arbitrary limit
					      * for now.
					      */
static struct cl_env_cache {
	rwlock_t		cec_guard;
	unsigned int		cec_count;
	struct list_head	cec_envs;
} *cl_envs = NULL;

struct cl_env {
	void	     *ce_magic;
	struct lu_env     ce_lu;
	struct lu_context ce_ses;

	/*
	 * Linkage into global list of all client environments. Used for
	 * garbage collection.
	 */
	struct list_head	ce_linkage;
	/*
	 *
	 */
	int	       ce_ref;
	/*
	 * Debugging field: address of the caller who made original
	 * allocation.
	 */
	void	     *ce_debug;
};

#define CL_ENV_INC(counter)
#define CL_ENV_DEC(counter)

static void cl_env_init0(struct cl_env *cle, void *debug)
{
	LASSERT(cle->ce_ref == 0);
	LASSERT(cle->ce_magic == &cl_env_init0);
	LASSERT(!cle->ce_debug);

	cle->ce_ref = 1;
	cle->ce_debug = debug;
	CL_ENV_INC(busy);
}

static struct lu_env *cl_env_new(__u32 ctx_tags, __u32 ses_tags, void *debug)
{
	struct lu_env *env;
	struct cl_env *cle;

	cle = kmem_cache_zalloc(cl_env_kmem, GFP_NOFS);
	if (cle) {
		int rc;

		INIT_LIST_HEAD(&cle->ce_linkage);
		cle->ce_magic = &cl_env_init0;
		env = &cle->ce_lu;
		rc = lu_env_init(env, ctx_tags | LCT_CL_THREAD);
		if (rc == 0) {
			rc = lu_context_init(&cle->ce_ses,
					     ses_tags | LCT_SESSION);
			if (rc == 0) {
				lu_context_enter(&cle->ce_ses);
				env->le_ses = &cle->ce_ses;
				cl_env_init0(cle, debug);
			} else {
				lu_env_fini(env);
			}
		}
		if (rc != 0) {
			kmem_cache_free(cl_env_kmem, cle);
			env = ERR_PTR(rc);
		} else {
			CL_ENV_INC(create);
			CL_ENV_INC(total);
		}
	} else {
		env = ERR_PTR(-ENOMEM);
	}
	return env;
}

static void cl_env_fini(struct cl_env *cle)
{
	CL_ENV_DEC(total);
	lu_context_fini(&cle->ce_lu.le_ctx);
	lu_context_fini(&cle->ce_ses);
	kmem_cache_free(cl_env_kmem, cle);
}

static struct lu_env *cl_env_obtain(void *debug)
{
	struct cl_env *cle;
	struct lu_env *env;
	int cpu = get_cpu();

	read_lock(&cl_envs[cpu].cec_guard);
	LASSERT(equi(cl_envs[cpu].cec_count == 0,
		     list_empty(&cl_envs[cpu].cec_envs)));
	if (cl_envs[cpu].cec_count > 0) {
		int rc;

		cle = container_of(cl_envs[cpu].cec_envs.next, struct cl_env,
				   ce_linkage);
		list_del_init(&cle->ce_linkage);
		cl_envs[cpu].cec_count--;
		read_unlock(&cl_envs[cpu].cec_guard);
		put_cpu();

		env = &cle->ce_lu;
		rc = lu_env_refill(env);
		if (rc == 0) {
			cl_env_init0(cle, debug);
			lu_context_enter(&env->le_ctx);
			lu_context_enter(&cle->ce_ses);
		} else {
			cl_env_fini(cle);
			env = ERR_PTR(rc);
		}
	} else {
		read_unlock(&cl_envs[cpu].cec_guard);
		put_cpu();
		env = cl_env_new(lu_context_tags_default,
				 lu_session_tags_default, debug);
	}
	return env;
}

static inline struct cl_env *cl_env_container(struct lu_env *env)
{
	return container_of(env, struct cl_env, ce_lu);
}

/**
 * Returns lu_env: if there already is an environment associated with the
 * current thread, it is returned, otherwise, new environment is allocated.
 *
 * Allocations are amortized through the global cache of environments.
 *
 * \param refcheck pointer to a counter used to detect environment leaks. In
 * the usual case cl_env_get() and cl_env_put() are called in the same lexical
 * scope and pointer to the same integer is passed as \a refcheck. This is
 * used to detect missed cl_env_put().
 *
 * \see cl_env_put()
 */
struct lu_env *cl_env_get(u16 *refcheck)
{
	struct lu_env *env;

	env = cl_env_obtain(__builtin_return_address(0));
	if (!IS_ERR(env)) {
		struct cl_env *cle;

		cle = cl_env_container(env);
		*refcheck = cle->ce_ref;
		CDEBUG(D_OTHER, "%d@%p\n", cle->ce_ref, cle);
	}
	return env;
}
EXPORT_SYMBOL(cl_env_get);

/**
 * Forces an allocation of a fresh environment with given tags.
 *
 * \see cl_env_get()
 */
struct lu_env *cl_env_alloc(u16 *refcheck, u32 tags)
{
	struct lu_env *env;

	env = cl_env_new(tags, tags, __builtin_return_address(0));
	if (!IS_ERR(env)) {
		struct cl_env *cle;

		cle = cl_env_container(env);
		*refcheck = cle->ce_ref;
		CDEBUG(D_OTHER, "%d@%p\n", cle->ce_ref, cle);
	}
	return env;
}
EXPORT_SYMBOL(cl_env_alloc);

static void cl_env_exit(struct cl_env *cle)
{
	lu_context_exit(&cle->ce_lu.le_ctx);
	lu_context_exit(&cle->ce_ses);
}

/**
 * Finalizes and frees a given number of cached environments. This is done to
 * (1) free some memory (not currently hooked into VM), or (2) release
 * references to modules.
 */
unsigned int cl_env_cache_purge(unsigned int nr)
{
	struct cl_env *cle;
	unsigned int i;

	for_each_possible_cpu(i) {
		write_lock(&cl_envs[i].cec_guard);
		for (; !list_empty(&cl_envs[i].cec_envs) && nr > 0; --nr) {
			cle = container_of(cl_envs[i].cec_envs.next,
					   struct cl_env, ce_linkage);
			list_del_init(&cle->ce_linkage);
			LASSERT(cl_envs[i].cec_count > 0);
			cl_envs[i].cec_count--;
			write_unlock(&cl_envs[i].cec_guard);

			cl_env_fini(cle);
			write_lock(&cl_envs[i].cec_guard);
		}
		LASSERT(equi(cl_envs[i].cec_count == 0,
			     list_empty(&cl_envs[i].cec_envs)));
		write_unlock(&cl_envs[i].cec_guard);
	}
	return nr;
}
EXPORT_SYMBOL(cl_env_cache_purge);

/**
 * Release an environment.
 *
 * Decrement \a env reference counter. When counter drops to 0, nothing in
 * this thread is using environment and it is returned to the allocation
 * cache, or freed straight away, if cache is large enough.
 */
void cl_env_put(struct lu_env *env, u16 *refcheck)
{
	struct cl_env *cle;

	cle = cl_env_container(env);

	LASSERT(cle->ce_ref > 0);
	LASSERT(ergo(refcheck, cle->ce_ref == *refcheck));

	CDEBUG(D_OTHER, "%d@%p\n", cle->ce_ref, cle);
	if (--cle->ce_ref == 0) {
		int cpu = get_cpu();

		CL_ENV_DEC(busy);
		cle->ce_debug = NULL;
		cl_env_exit(cle);
		/*
		 * Don't bother to take a lock here.
		 *
		 * Return environment to the cache only when it was allocated
		 * with the standard tags.
		 */
		if (cl_envs[cpu].cec_count < cl_envs_cached_max &&
		    (env->le_ctx.lc_tags & ~LCT_HAS_EXIT) == LCT_CL_THREAD &&
		    (env->le_ses->lc_tags & ~LCT_HAS_EXIT) == LCT_SESSION) {
			read_lock(&cl_envs[cpu].cec_guard);
			list_add(&cle->ce_linkage, &cl_envs[cpu].cec_envs);
			cl_envs[cpu].cec_count++;
			read_unlock(&cl_envs[cpu].cec_guard);
		} else {
			cl_env_fini(cle);
		}
		put_cpu();
	}
}
EXPORT_SYMBOL(cl_env_put);

/**
 * Converts struct ost_lvb to struct cl_attr.
 *
 * \see cl_attr2lvb
 */
void cl_lvb2attr(struct cl_attr *attr, const struct ost_lvb *lvb)
{
	attr->cat_size   = lvb->lvb_size;
	attr->cat_mtime  = lvb->lvb_mtime;
	attr->cat_atime  = lvb->lvb_atime;
	attr->cat_ctime  = lvb->lvb_ctime;
	attr->cat_blocks = lvb->lvb_blocks;
}
EXPORT_SYMBOL(cl_lvb2attr);

static struct cl_env cl_env_percpu[NR_CPUS];

static int cl_env_percpu_init(void)
{
	struct cl_env *cle;
	int tags = LCT_REMEMBER | LCT_NOREF;
	int i, j;
	int rc = 0;

	for_each_possible_cpu(i) {
		struct lu_env *env;

		rwlock_init(&cl_envs[i].cec_guard);
		INIT_LIST_HEAD(&cl_envs[i].cec_envs);
		cl_envs[i].cec_count = 0;

		cle = &cl_env_percpu[i];
		env = &cle->ce_lu;

		INIT_LIST_HEAD(&cle->ce_linkage);
		cle->ce_magic = &cl_env_init0;
		rc = lu_env_init(env, LCT_CL_THREAD | tags);
		if (rc == 0) {
			rc = lu_context_init(&cle->ce_ses, LCT_SESSION | tags);
			if (rc == 0) {
				lu_context_enter(&cle->ce_ses);
				env->le_ses = &cle->ce_ses;
			} else {
				lu_env_fini(env);
			}
		}
		if (rc != 0)
			break;
	}
	if (rc != 0) {
		/* Indices 0 to i (excluding i) were correctly initialized,
		 * thus we must uninitialize up to i, the rest are undefined.
		 */
		for (j = 0; j < i; j++) {
			cle = &cl_env_percpu[j];
			lu_context_exit(&cle->ce_ses);
			lu_context_fini(&cle->ce_ses);
			lu_env_fini(&cle->ce_lu);
		}
	}

	return rc;
}

static void cl_env_percpu_fini(void)
{
	int i;

	for_each_possible_cpu(i) {
		struct cl_env *cle = &cl_env_percpu[i];

		lu_context_exit(&cle->ce_ses);
		lu_context_fini(&cle->ce_ses);
		lu_env_fini(&cle->ce_lu);
	}
}

static void cl_env_percpu_refill(void)
{
	int i;

	for_each_possible_cpu(i)
		lu_env_refill(&cl_env_percpu[i].ce_lu);
}

void cl_env_percpu_put(struct lu_env *env)
{
	struct cl_env *cle;
	int cpu;

	cpu = smp_processor_id();
	cle = cl_env_container(env);
	LASSERT(cle == &cl_env_percpu[cpu]);

	cle->ce_ref--;
	LASSERT(cle->ce_ref == 0);

	CL_ENV_DEC(busy);
	cle->ce_debug = NULL;

	put_cpu();
}
EXPORT_SYMBOL(cl_env_percpu_put);

struct lu_env *cl_env_percpu_get(void)
{
	struct cl_env *cle;

	cle = &cl_env_percpu[get_cpu()];
	cl_env_init0(cle, __builtin_return_address(0));

	return &cle->ce_lu;
}
EXPORT_SYMBOL(cl_env_percpu_get);

/*****************************************************************************
 *
 * Temporary prototype thing: mirror obd-devices into cl devices.
 *
 */

struct cl_device *cl_type_setup(const struct lu_env *env, struct lu_site *site,
				struct lu_device_type *ldt,
				struct lu_device *next)
{
	const char       *typename;
	struct lu_device *d;

	typename = ldt->ldt_name;
	d = ldt->ldt_ops->ldto_device_alloc(env, ldt, NULL);
	if (!IS_ERR(d)) {
		int rc;

		if (site)
			d->ld_site = site;
		rc = ldt->ldt_ops->ldto_device_init(env, d, typename, next);
		if (rc == 0) {
			lu_device_get(d);
			lu_ref_add(&d->ld_reference,
				   "lu-stack", &lu_site_init);
		} else {
			ldt->ldt_ops->ldto_device_free(env, d);
			CERROR("can't init device '%s', %d\n", typename, rc);
			d = ERR_PTR(rc);
		}
	} else {
		CERROR("Cannot allocate device: '%s'\n", typename);
	}
	return lu2cl_dev(d);
}
EXPORT_SYMBOL(cl_type_setup);

/**
 * Finalize device stack by calling lu_stack_fini().
 */
void cl_stack_fini(const struct lu_env *env, struct cl_device *cl)
{
	lu_stack_fini(env, cl2lu_dev(cl));
}
EXPORT_SYMBOL(cl_stack_fini);

static struct lu_context_key cl_key;

struct cl_thread_info *cl_env_info(const struct lu_env *env)
{
	return lu_context_key_get(&env->le_ctx, &cl_key);
}

/* defines cl0_key_{init,fini}() */
LU_KEY_INIT_FINI(cl0, struct cl_thread_info);

static void *cl_key_init(const struct lu_context *ctx,
			 struct lu_context_key *key)
{
	return cl0_key_init(ctx, key);
}

static void cl_key_fini(const struct lu_context *ctx,
			struct lu_context_key *key, void *data)
{
	cl0_key_fini(ctx, key, data);
}

static struct lu_context_key cl_key = {
	.lct_tags = LCT_CL_THREAD,
	.lct_init = cl_key_init,
	.lct_fini = cl_key_fini,
};

static struct lu_kmem_descr cl_object_caches[] = {
	{
		.ckd_cache = &cl_env_kmem,
		.ckd_name  = "cl_env_kmem",
		.ckd_size  = sizeof(struct cl_env)
	},
	{
		.ckd_cache = NULL
	}
};

/**
 * Global initialization of cl-data. Create kmem caches, register
 * lu_context_key's, etc.
 *
 * \see cl_global_fini()
 */
int cl_global_init(void)
{
	int result;

	cl_envs = kzalloc(sizeof(*cl_envs) * num_possible_cpus(), GFP_KERNEL);
	if (!cl_envs) {
		result = -ENOMEM;
		goto out;
	}

	result = lu_kmem_init(cl_object_caches);
	if (result)
		goto out_envs;

	LU_CONTEXT_KEY_INIT(&cl_key);
	result = lu_context_key_register(&cl_key);
	if (result)
		goto out_kmem;

	result = cl_env_percpu_init();
	if (result)
		/* no cl_env_percpu_fini on error */
		goto out_keys;

	return 0;

out_keys:
	lu_context_key_degister(&cl_key);
out_kmem:
	lu_kmem_fini(cl_object_caches);
out_envs:
	kfree(cl_envs);
out:
	return result;
}

/**
 * Finalization of global cl-data. Dual to cl_global_init().
 */
void cl_global_fini(void)
{
	cl_env_percpu_fini();
	lu_context_key_degister(&cl_key);
	lu_kmem_fini(cl_object_caches);
	kfree(cl_envs);
}