blob: 36afe99e4f93e1e62386700f57c9739c55ac5b33 [file] [log] [blame]
David Teiglande7fd4172006-01-18 09:30:29 +00001/******************************************************************************
2*******************************************************************************
3**
4** Copyright (C) 2005 Red Hat, Inc. All rights reserved.
5**
6** This copyrighted material is made available to anyone wishing to use,
7** modify, copy, or redistribute it subject to the terms and conditions
8** of the GNU General Public License v.2.
9**
10*******************************************************************************
11******************************************************************************/
12
13#include "dlm_internal.h"
14#include "member.h"
15#include "lock.h"
16#include "dir.h"
17#include "config.h"
18#include "requestqueue.h"
19
20struct rq_entry {
21 struct list_head list;
22 int nodeid;
23 char request[1];
24};
25
26/*
27 * Requests received while the lockspace is in recovery get added to the
28 * request queue and processed when recovery is complete. This happens when
29 * the lockspace is suspended on some nodes before it is on others, or the
30 * lockspace is enabled on some while still suspended on others.
31 */
32
33void dlm_add_requestqueue(struct dlm_ls *ls, int nodeid, struct dlm_header *hd)
34{
35 struct rq_entry *e;
36 int length = hd->h_length;
37
38 if (dlm_is_removed(ls, nodeid))
39 return;
40
41 e = kmalloc(sizeof(struct rq_entry) + length, GFP_KERNEL);
42 if (!e) {
43 log_print("dlm_add_requestqueue: out of memory\n");
44 return;
45 }
46
47 e->nodeid = nodeid;
48 memcpy(e->request, hd, length);
49
50 down(&ls->ls_requestqueue_lock);
51 list_add_tail(&e->list, &ls->ls_requestqueue);
52 up(&ls->ls_requestqueue_lock);
53}
54
55int dlm_process_requestqueue(struct dlm_ls *ls)
56{
57 struct rq_entry *e;
58 struct dlm_header *hd;
59 int error = 0;
60
61 down(&ls->ls_requestqueue_lock);
62
63 for (;;) {
64 if (list_empty(&ls->ls_requestqueue)) {
65 up(&ls->ls_requestqueue_lock);
66 error = 0;
67 break;
68 }
69 e = list_entry(ls->ls_requestqueue.next, struct rq_entry, list);
70 up(&ls->ls_requestqueue_lock);
71
72 hd = (struct dlm_header *) e->request;
73 error = dlm_receive_message(hd, e->nodeid, TRUE);
74
75 if (error == -EINTR) {
76 /* entry is left on requestqueue */
77 log_debug(ls, "process_requestqueue abort eintr");
78 break;
79 }
80
81 down(&ls->ls_requestqueue_lock);
82 list_del(&e->list);
83 kfree(e);
84
85 if (dlm_locking_stopped(ls)) {
86 log_debug(ls, "process_requestqueue abort running");
87 up(&ls->ls_requestqueue_lock);
88 error = -EINTR;
89 break;
90 }
91 schedule();
92 }
93
94 return error;
95}
96
97/*
98 * After recovery is done, locking is resumed and dlm_recoverd takes all the
99 * saved requests and processes them as they would have been by dlm_recvd. At
100 * the same time, dlm_recvd will start receiving new requests from remote
101 * nodes. We want to delay dlm_recvd processing new requests until
102 * dlm_recoverd has finished processing the old saved requests.
103 */
104
105void dlm_wait_requestqueue(struct dlm_ls *ls)
106{
107 for (;;) {
108 down(&ls->ls_requestqueue_lock);
109 if (list_empty(&ls->ls_requestqueue))
110 break;
111 if (dlm_locking_stopped(ls))
112 break;
113 up(&ls->ls_requestqueue_lock);
114 schedule();
115 }
116 up(&ls->ls_requestqueue_lock);
117}
118
119static int purge_request(struct dlm_ls *ls, struct dlm_message *ms, int nodeid)
120{
121 uint32_t type = ms->m_type;
122
123 if (dlm_is_removed(ls, nodeid))
124 return 1;
125
126 /* directory operations are always purged because the directory is
127 always rebuilt during recovery and the lookups resent */
128
129 if (type == DLM_MSG_REMOVE ||
130 type == DLM_MSG_LOOKUP ||
131 type == DLM_MSG_LOOKUP_REPLY)
132 return 1;
133
134 if (!dlm_no_directory(ls))
135 return 0;
136
137 /* with no directory, the master is likely to change as a part of
138 recovery; requests to/from the defunct master need to be purged */
139
140 switch (type) {
141 case DLM_MSG_REQUEST:
142 case DLM_MSG_CONVERT:
143 case DLM_MSG_UNLOCK:
144 case DLM_MSG_CANCEL:
145 /* we're no longer the master of this resource, the sender
146 will resend to the new master (see waiter_needs_recovery) */
147
148 if (dlm_hash2nodeid(ls, ms->m_hash) != dlm_our_nodeid())
149 return 1;
150 break;
151
152 case DLM_MSG_REQUEST_REPLY:
153 case DLM_MSG_CONVERT_REPLY:
154 case DLM_MSG_UNLOCK_REPLY:
155 case DLM_MSG_CANCEL_REPLY:
156 case DLM_MSG_GRANT:
157 /* this reply is from the former master of the resource,
158 we'll resend to the new master if needed */
159
160 if (dlm_hash2nodeid(ls, ms->m_hash) != nodeid)
161 return 1;
162 break;
163 }
164
165 return 0;
166}
167
168void dlm_purge_requestqueue(struct dlm_ls *ls)
169{
170 struct dlm_message *ms;
171 struct rq_entry *e, *safe;
172
173 down(&ls->ls_requestqueue_lock);
174 list_for_each_entry_safe(e, safe, &ls->ls_requestqueue, list) {
175 ms = (struct dlm_message *) e->request;
176
177 if (purge_request(ls, ms, e->nodeid)) {
178 list_del(&e->list);
179 kfree(e);
180 }
181 }
182 up(&ls->ls_requestqueue_lock);
183}
184