Lustre Repository Browser

Viewing: fld_request.c

// SPDX-License-Identifier: GPL-2.0

/*
 * Copyright (c) 2007, 2010, Oracle and/or its affiliates. All rights reserved.
 * Use is subject to license terms.
 *
 * Copyright (c) 2011, 2017, Intel Corporation.
 */

/*
 * This file is part of Lustre, http://www.lustre.org/
 *
 * FLD (Fids Location Database)
 *
 * Author: Yury Umanets <umka@clusterfs.com>
 */

#define DEBUG_SUBSYSTEM S_FLD

#include <linux/module.h>
#include <linux/math64.h>
#include <linux/delay.h>

#include <obd.h>
#include <obd_class.h>
#include <obd_support.h>
#include <lprocfs_status.h>
#include <lustre_req_layout.h>
#include <lustre_fld.h>
#include <lustre_mdc.h>
#include "fld_internal.h"

static int fld_rrb_hash(struct lu_client_fld *fld, u64 seq)
{
	LASSERT(fld->lcf_count > 0);
	return do_div(seq, fld->lcf_count);
}

static struct lu_fld_target *
fld_rrb_scan(struct lu_client_fld *fld, u64 seq)
{
	struct lu_fld_target *target;
	int hash;

	ENTRY;

	/*
	 * Because almost all of special sequence located in MDT0,
	 * it should go to index 0 directly, instead of calculating
	 * hash again, and also if other MDTs is not being connected,
	 * the fld lookup requests(for seq on MDT0) should not be
	 * blocked because of other MDTs
	 */
	if (fid_seq_is_norm(seq))
		hash = fld_rrb_hash(fld, seq);
	else
		hash = 0;

again:
	list_for_each_entry(target, &fld->lcf_targets, ft_chain) {
		if (target->ft_idx == hash)
			RETURN(target);
	}

	if (hash != 0) {
		/*
		 * It is possible the remote target(MDT) are not connected to
		 * with client yet, so we will refer this to MDT0, which should
		 * be connected during mount
		 */
		hash = 0;
		goto again;
	}

	CERROR("%s: Can't find target by hash %d (seq %#llx). Targets (%d):\n",
	       fld->lcf_name, hash, seq, fld->lcf_count);

	list_for_each_entry(target, &fld->lcf_targets, ft_chain) {
		const char *srv_name = target->ft_srv != NULL  ?
			target->ft_srv->lsf_name : "<null>";
		const char *exp_name = target->ft_exp != NULL ?
			(char *)target->ft_exp->exp_obd->obd_uuid.uuid :
			"<null>";

		CERROR("  exp: 0x%p (%s), srv: 0x%p (%s), idx: %llu\n",
		       target->ft_exp, exp_name, target->ft_srv,
		       srv_name, target->ft_idx);
	}

	/*
	 * If target is not found, there is logical error anyway, so here is
	 * LBUG() to catch this situation.
	 */
	LBUG();
	RETURN(NULL);
}

struct lu_fld_hash fld_hash[] = {
	{
		.fh_name = "RRB",
		.fh_hash_func = fld_rrb_hash,
		.fh_scan_func = fld_rrb_scan
	},
	{
		NULL,
	}
};

static struct lu_fld_target *
fld_client_get_target(struct lu_client_fld *fld, u64 seq)
{
	struct lu_fld_target *target;

	ENTRY;

	LASSERT(fld->lcf_hash != NULL);

	spin_lock(&fld->lcf_lock);
	target = fld->lcf_hash->fh_scan_func(fld, seq);
	spin_unlock(&fld->lcf_lock);

	if (target) {
		CDEBUG(D_INFO, "%s: Found target (idx %llu) by seq %#llx\n",
		       fld->lcf_name, target->ft_idx, seq);
	}

	RETURN(target);
}

/*
 * Add export to FLD. This is usually done by CMM and LMV as they are main users
 * of FLD module.
 */
int fld_client_add_target(struct lu_client_fld *fld,
			  struct lu_fld_target *tar)
{
	const char *name;
	struct lu_fld_target *target, *tmp;

	ENTRY;

	LASSERT(tar != NULL);
	name = fld_target_name(tar);
	LASSERT(name != NULL);
	LASSERT(tar->ft_srv != NULL || tar->ft_exp != NULL);

	CDEBUG(D_INFO, "%s: Adding target %s (idx %llu)\n", fld->lcf_name,
	       name, tar->ft_idx);

	OBD_ALLOC_PTR(target);
	if (!target)
		RETURN(-ENOMEM);

	spin_lock(&fld->lcf_lock);
	list_for_each_entry(tmp, &fld->lcf_targets, ft_chain) {
		if (tmp->ft_idx == tar->ft_idx) {
			spin_unlock(&fld->lcf_lock);
			OBD_FREE_PTR(target);
			CERROR("Target %s exists in FLD and known as %s:#%llu\n",
			       name, fld_target_name(tmp), tmp->ft_idx);
			RETURN(-EEXIST);
		}
	}

	target->ft_exp = tar->ft_exp;
	if (target->ft_exp)
		class_export_get(target->ft_exp);
	target->ft_srv = tar->ft_srv;
	target->ft_idx = tar->ft_idx;

	list_add_tail(&target->ft_chain, &fld->lcf_targets);

	fld->lcf_count++;
	spin_unlock(&fld->lcf_lock);

	RETURN(0);
}
EXPORT_SYMBOL(fld_client_add_target);

/* Remove export from FLD */
int fld_client_del_target(struct lu_client_fld *fld, u64 idx)
{
	struct lu_fld_target *target, *tmp;

	ENTRY;

	spin_lock(&fld->lcf_lock);
	list_for_each_entry_safe(target, tmp, &fld->lcf_targets, ft_chain) {
		if (target->ft_idx == idx) {
			fld->lcf_count--;
			list_del(&target->ft_chain);
			spin_unlock(&fld->lcf_lock);

			if (target->ft_exp)
				class_export_put(target->ft_exp);

			OBD_FREE_PTR(target);
			RETURN(0);
		}
	}
	spin_unlock(&fld->lcf_lock);
	RETURN(-ENOENT);
}
EXPORT_SYMBOL(fld_client_del_target);

struct dentry *fld_debugfs_dir;

static void fld_client_debugfs_init(struct lu_client_fld *fld)
{
	ENTRY;
	fld->lcf_debugfs_entry = debugfs_create_dir(fld->lcf_name,
						   fld_debugfs_dir);
	ldebugfs_add_vars(fld->lcf_debugfs_entry,
			  fld_client_debugfs_list,
			  fld);
}

void fld_client_debugfs_fini(struct lu_client_fld *fld)
{
	debugfs_remove_recursive(fld->lcf_debugfs_entry);
}
EXPORT_SYMBOL(fld_client_debugfs_fini);

static inline int hash_is_sane(int hash)
{
	return (hash >= 0 && hash < ARRAY_SIZE(fld_hash));
}

int fld_client_init(struct lu_client_fld *fld,
		    const char *prefix, int hash)
{
	int cache_size, cache_threshold;
	int rc = 0;

	ENTRY;
	snprintf(fld->lcf_name, sizeof(fld->lcf_name),
		 "cli-%s", prefix);

	if (!hash_is_sane(hash)) {
		CERROR("%s: Wrong hash function %#x\n",
		       fld->lcf_name, hash);
		RETURN(-EINVAL);
	}

	fld->lcf_count = 0;
	spin_lock_init(&fld->lcf_lock);
	fld->lcf_hash = &fld_hash[hash];
	INIT_LIST_HEAD(&fld->lcf_targets);

	cache_size = FLD_CLIENT_CACHE_SIZE /
		sizeof(struct fld_cache_entry);

	cache_threshold = cache_size *
		FLD_CLIENT_CACHE_THRESHOLD / 100;

	fld->lcf_cache = fld_cache_init(fld->lcf_name,
					cache_size, cache_threshold);
	if (IS_ERR(fld->lcf_cache)) {
		rc = PTR_ERR(fld->lcf_cache);
		fld->lcf_cache = NULL;
		GOTO(out, rc);
	}

	fld_client_debugfs_init(fld);
	EXIT;
out:
	if (rc)
		fld_client_fini(fld);
	else
		CDEBUG(D_INFO, "%s: Using \"%s\" hash\n",
		       fld->lcf_name, fld->lcf_hash->fh_name);
	return rc;
}
EXPORT_SYMBOL(fld_client_init);

void fld_client_fini(struct lu_client_fld *fld)
{
	struct lu_fld_target *target, *tmp;

	ENTRY;

	spin_lock(&fld->lcf_lock);
	list_for_each_entry_safe(target, tmp, &fld->lcf_targets, ft_chain) {
		fld->lcf_count--;
		list_del(&target->ft_chain);
		if (target->ft_exp)
			class_export_put(target->ft_exp);
		OBD_FREE_PTR(target);
	}
	spin_unlock(&fld->lcf_lock);

	if (fld->lcf_cache) {
		if (!IS_ERR(fld->lcf_cache))
			fld_cache_fini(fld->lcf_cache);
		fld->lcf_cache = NULL;
	}

	EXIT;
}
EXPORT_SYMBOL(fld_client_fini);

int fld_client_rpc(struct obd_export *exp,
		   struct lu_seq_range *range, u32 fld_op,
		   struct ptlrpc_request **reqp)
{
	struct ptlrpc_request *req = NULL;
	struct lu_seq_range *prange;
	u32 *op;
	int rc = 0;
	struct obd_import *imp;

	ENTRY;

	LASSERT(exp != NULL);

	imp = class_exp2cliimp(exp);
again:
	switch (fld_op) {
	case FLD_QUERY:
		req = ptlrpc_request_alloc_pack(imp, &RQF_FLD_QUERY,
						LUSTRE_MDS_VERSION, FLD_QUERY);
		if (IS_ERR(req))
			RETURN(PTR_ERR(req));

		/*
		 * XXX: only needed when talking to old server(< 2.6), it should
		 * be removed when < 2.6 server is not supported
		 */
		op = req_capsule_client_get(&req->rq_pill, &RMF_FLD_OPC);
		*op = FLD_LOOKUP;

		/*
		 * For MDS_MDS seq lookup, it will always use LWP connection,
		 * but LWP will be evicted after restart, so cause the error.
		 * so we will set no_delay for seq lookup request, once the
		 * request fails because of the eviction. always retry here
		 */
		if (imp->imp_connect_flags_orig & OBD_CONNECT_MDS_MDS) {
			req->rq_allow_replay = 1;
			req->rq_no_delay = 1;
		}
		break;
	case FLD_READ:
		req = ptlrpc_request_alloc_pack(imp, &RQF_FLD_READ,
						LUSTRE_MDS_VERSION, FLD_READ);
		if (IS_ERR(req))
			RETURN(PTR_ERR(req));

		req_capsule_set_size(&req->rq_pill, &RMF_GENERIC_DATA,
				     RCL_SERVER, PAGE_SIZE);
		break;
	default:
		rc = -EINVAL;
		break;
	}

	if (rc != 0)
		RETURN(rc);

	prange = req_capsule_client_get(&req->rq_pill, &RMF_FLD_MDFLD);
	*prange = *range;
	ptlrpc_request_set_replen(req);
	req->rq_request_portal = FLD_REQUEST_PORTAL;
	req->rq_reply_portal = MDC_REPLY_PORTAL;
	ptlrpc_at_set_req_timeout(req);

	if (CFS_FAIL_CHECK(OBD_FAIL_FLD_QUERY_REQ) && req->rq_no_delay) {
		/* the same error returned by ptlrpc_import_delay_req */
		rc = -EAGAIN;
		req->rq_status = rc;
	} else {
		rc = ptlrpc_queue_wait(req);
	}

	if (rc == -ENOENT) {
		/* Don't loop forever on non-existing FID sequences. */
		GOTO(out_req, rc);
	}

	if (rc != 0) {
		if (imp->imp_state != LUSTRE_IMP_CLOSED &&
		    !test_bit(IMPF_DEACTIVE, imp->imp_flags) &&
		    imp->imp_connect_flags_orig & OBD_CONNECT_MDS_MDS &&
		    OCD_HAS_FLAG(&imp->imp_connect_data, LIGHTWEIGHT) &&
		    rc != -ENOTSUPP) {
			/* LWP is not replayable, retry after a while */
			rc = -EAGAIN;
		}
		if (rc == -EAGAIN) {
			if (msleep_interruptible(2 * MSEC_PER_SEC))
				GOTO(out_req, rc = -EINTR);
			ptlrpc_req_put(req);
			rc = 0;
			goto again;
		}
		GOTO(out_req, rc);
	}

	if (fld_op == FLD_QUERY) {
		prange = req_capsule_server_get(&req->rq_pill,
						&RMF_FLD_MDFLD);
		if (!prange)
			GOTO(out_req, rc = -EFAULT);
		*range = *prange;
	}

	EXIT;
out_req:
	if (rc != 0 || !reqp) {
		ptlrpc_req_put(req);
		req = NULL;
	}

	if (reqp)
		*reqp = req;

	return rc;
}

int fld_client_lookup(struct lu_client_fld *fld, u64 seq, u32 flags,
		      const struct lu_env *env, struct lu_seq_range *res)
{
	struct lu_fld_target *target;
	struct lu_fld_target *origin;
	int rc;

	ENTRY;

	LASSERT(fld != NULL);
	LASSERT(res != NULL);

	rc = fld_cache_lookup(fld->lcf_cache, seq, res);
	if (rc == 0)
		RETURN(0);

	/* Can not find it in the cache */
	target = fld_client_get_target(fld, seq);
	LASSERT(target != NULL);
	origin = target;

again:
	CDEBUG(D_INFO, "%s: Lookup fld entry (seq: %#llx) on target %s (idx %llu)\n",
	       fld->lcf_name, seq, fld_target_name(target), target->ft_idx);

	res->lsr_start = seq;
	fld_range_set_type(res, flags);

#ifdef CONFIG_LUSTRE_FS_SERVER
	if (target->ft_srv) {
		LASSERT(env != NULL);
		rc = fld_server_lookup(env, target->ft_srv, seq, res);
	} else
#endif /* CONFIG_LUSTRE_FS_SERVER */
	{
		rc = fld_client_rpc(target->ft_exp, res, FLD_QUERY, NULL);
	}

	if (rc == -ESHUTDOWN) {
		/*
		 * If fld lookup failed because the target has been shutdown,
		 * then try next target in the list, until trying all targets
		 * or fld lookup succeeds
		 */
		spin_lock(&fld->lcf_lock);
		/*
		 * If the next entry in the list is the head of the list,
		 * move to the next entry after the head and retrieve
		 * the target. Else retreive the next target entry.
		 */
		if (target->ft_chain.next == &fld->lcf_targets)
			target = list_first_entry(target->ft_chain.next,
						  struct lu_fld_target,
						  ft_chain);
		else
			target = list_first_entry(&target->ft_chain,
						 struct lu_fld_target,
						 ft_chain);
		spin_unlock(&fld->lcf_lock);
		if (target != origin)
			goto again;
	}
	if (rc == 0)
		fld_cache_insert(fld->lcf_cache, res);

	RETURN(rc);
}
EXPORT_SYMBOL(fld_client_lookup);

void fld_client_flush(struct lu_client_fld *fld)
{
	fld_cache_flush(fld->lcf_cache);
}

static int __init fld_init(void)
{
	int rc;

	rc = libcfs_setup();
	if (rc)
		return rc;

#ifdef CONFIG_LUSTRE_FS_SERVER
	rc = fld_server_mod_init();
	if (rc)
		return rc;
#endif /* CONFIG_LUSTRE_FS_SERVER */

	fld_debugfs_dir = debugfs_create_dir(LUSTRE_FLD_NAME,
					     debugfs_lustre_root);
	return PTR_ERR_OR_ZERO(fld_debugfs_dir);
}

static void __exit fld_exit(void)
{
#ifdef CONFIG_LUSTRE_FS_SERVER
	fld_server_mod_exit();
#endif /* CONFIG_LUSTRE_FS_SERVER */
	debugfs_remove_recursive(fld_debugfs_dir);
}

MODULE_AUTHOR("OpenSFS, Inc. <http://www.lustre.org/>");
MODULE_DESCRIPTION("Lustre FID Location Database");
MODULE_VERSION(LUSTRE_VERSION_STRING);
MODULE_LICENSE("GPL");

late_initcall_sync(fld_init);
module_exit(fld_exit);