summaryrefslogtreecommitdiff
path: root/fs/dlm/requestqueue.c
blob: 892d6ca21e7424f84227718a0ae8064a32301958 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
// SPDX-License-Identifier: GPL-2.0-only
/******************************************************************************
*******************************************************************************
**
**  Copyright (C) 2005-2007 Red Hat, Inc.  All rights reserved.
**
**
*******************************************************************************
******************************************************************************/

#include "dlm_internal.h"
#include "member.h"
#include "lock.h"
#include "dir.h"
#include "config.h"
#include "requestqueue.h"
#include "util.h"

struct rq_entry {
	struct list_head list;
	uint32_t recover_seq;
	int nodeid;
	struct dlm_message request;
};

/*
 * Requests received while the lockspace is in recovery get added to the
 * request queue and processed when recovery is complete.  This happens when
 * the lockspace is suspended on some nodes before it is on others, or the
 * lockspace is enabled on some while still suspended on others.
 */

void dlm_add_requestqueue(struct dlm_ls *ls, int nodeid,
			  const struct dlm_message *ms)
{
	struct rq_entry *e;
	int length = le16_to_cpu(ms->m_header.h_length) -
		sizeof(struct dlm_message);

	e = kmalloc(sizeof(struct rq_entry) + length, GFP_NOFS);
	if (!e) {
		log_print("dlm_add_requestqueue: out of memory len %d", length);
		return;
	}

	e->recover_seq = ls->ls_recover_seq & 0xFFFFFFFF;
	e->nodeid = nodeid;
	memcpy(&e->request, ms, sizeof(*ms));
	memcpy(&e->request.m_extra, ms->m_extra, length);

	atomic_inc(&ls->ls_requestqueue_cnt);
	mutex_lock(&ls->ls_requestqueue_mutex);
	list_add_tail(&e->list, &ls->ls_requestqueue);
	mutex_unlock(&ls->ls_requestqueue_mutex);
}

/*
 * Called by dlm_recoverd to process normal messages saved while recovery was
 * happening.  Normal locking has been enabled before this is called.  dlm_recv
 * upon receiving a message, will wait for all saved messages to be drained
 * here before processing the message it got.  If a new dlm_ls_stop() arrives
 * while we're processing these saved messages, it may block trying to suspend
 * dlm_recv if dlm_recv is waiting for us in dlm_wait_requestqueue.  In that
 * case, we don't abort since locking_stopped is still 0.  If dlm_recv is not
 * waiting for us, then this processing may be aborted due to locking_stopped.
 */

int dlm_process_requestqueue(struct dlm_ls *ls)
{
	struct rq_entry *e;
	struct dlm_message *ms;
	int error = 0;

	mutex_lock(&ls->ls_requestqueue_mutex);

	for (;;) {
		if (list_empty(&ls->ls_requestqueue)) {
			mutex_unlock(&ls->ls_requestqueue_mutex);
			error = 0;
			break;
		}
		e = list_entry(ls->ls_requestqueue.next, struct rq_entry, list);
		mutex_unlock(&ls->ls_requestqueue_mutex);

		ms = &e->request;

		log_limit(ls, "dlm_process_requestqueue msg %d from %d "
			  "lkid %x remid %x result %d seq %u",
			  le32_to_cpu(ms->m_type),
			  le32_to_cpu(ms->m_header.h_nodeid),
			  le32_to_cpu(ms->m_lkid), le32_to_cpu(ms->m_remid),
			  from_dlm_errno(le32_to_cpu(ms->m_result)),
			  e->recover_seq);

		dlm_receive_message_saved(ls, &e->request, e->recover_seq);

		mutex_lock(&ls->ls_requestqueue_mutex);
		list_del(&e->list);
		if (atomic_dec_and_test(&ls->ls_requestqueue_cnt))
			wake_up(&ls->ls_requestqueue_wait);
		kfree(e);

		if (dlm_locking_stopped(ls)) {
			log_debug(ls, "process_requestqueue abort running");
			mutex_unlock(&ls->ls_requestqueue_mutex);
			error = -EINTR;
			break;
		}
		schedule();
	}

	return error;
}

/*
 * After recovery is done, locking is resumed and dlm_recoverd takes all the
 * saved requests and processes them as they would have been by dlm_recv.  At
 * the same time, dlm_recv will start receiving new requests from remote nodes.
 * We want to delay dlm_recv processing new requests until dlm_recoverd has
 * finished processing the old saved requests.  We don't check for locking
 * stopped here because dlm_ls_stop won't stop locking until it's suspended us
 * (dlm_recv).
 */

void dlm_wait_requestqueue(struct dlm_ls *ls)
{
	wait_event(ls->ls_requestqueue_wait,
		   atomic_read(&ls->ls_requestqueue_cnt) == 0);
}

static int purge_request(struct dlm_ls *ls, struct dlm_message *ms, int nodeid)
{
	__le32 type = ms->m_type;

	/* the ls is being cleaned up and freed by release_lockspace */
	if (!atomic_read(&ls->ls_count))
		return 1;

	if (dlm_is_removed(ls, nodeid))
		return 1;

	/* directory operations are always purged because the directory is
	   always rebuilt during recovery and the lookups resent */

	if (type == cpu_to_le32(DLM_MSG_REMOVE) ||
	    type == cpu_to_le32(DLM_MSG_LOOKUP) ||
	    type == cpu_to_le32(DLM_MSG_LOOKUP_REPLY))
		return 1;

	if (!dlm_no_directory(ls))
		return 0;

	return 1;
}

void dlm_purge_requestqueue(struct dlm_ls *ls)
{
	struct dlm_message *ms;
	struct rq_entry *e, *safe;

	mutex_lock(&ls->ls_requestqueue_mutex);
	list_for_each_entry_safe(e, safe, &ls->ls_requestqueue, list) {
		ms =  &e->request;

		if (purge_request(ls, ms, e->nodeid)) {
			list_del(&e->list);
			if (atomic_dec_and_test(&ls->ls_requestqueue_cnt))
				wake_up(&ls->ls_requestqueue_wait);
			kfree(e);
		}
	}
	mutex_unlock(&ls->ls_requestqueue_mutex);
}