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
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
|
/*
* Berkeley Lab Checkpoint/Restart (BLCR) for Linux is Copyright (c)
* 2003, The Regents of the University of California, through Lawrence
* Berkeley National Laboratory (subject to receipt of any required
* approvals from the U.S. Dept. of Energy). All rights reserved.
*
* Portions may be copyrighted by others, as may be noted in specific
* copyright notices within specific files.
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* 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 for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
*
* $Id: cr_trigger.c,v 1.39 2008/12/05 23:15:19 phargrov Exp $
*/
#include "cr_module.h"
//
// Private functions
//
// do_trigger(cr_task, proc_req)
//
// Sends a trigger (currently just signal 63 or 64) to the given task.
// If the task is STOPPED, we also wake it so it can receive the
// signal.
// Must be called with tasklist_lock held for reading to prevent task from exiting.
//
// XXX: Should (probably) never fail, but could return -EAGAIN if the
// kernel can't queue anymore RT signals at this time.
// XXX: When we finally move away from the signal, the trigger will
// need to invoke the cr_task->handler, or the default trigger if NULL.
static int
do_trigger(cr_task_t *cr_task, cr_chkpt_proc_req_t *proc_req)
{
struct task_struct *task = cr_task->task;
struct siginfo info;
int retval = 0;
if (cri_task_dead(task)) {
CR_KTRACE_LOW_LVL("not triggering dead task pid %d (%s)", task->pid, task->comm);
goto out;
}
retval = proc_req->ctrl_fd;
if (retval < 0) {
retval = proc_req->tmp_fd;
if (retval < 0) {
/* Force an instance of the ctrl fd into the target's file table
* (exactly once per process regardless of thread count).
* XXX: Ideally we'd be smarter about this:
* STOP other tasks before doing this to avoid racing against
* anything they might be doing with their own files.
*/
retval = cr_dup_other(task->files, proc_req->req->ctrl_file);
if (retval < 0) goto out;
proc_req->tmp_fd = retval;
}
}
CR_ASSERT_STEP_EQ(cr_task, 0);
memset(&info, 0, sizeof(info)); /* security */
info.si_signo = CR_SIGNUM;
info.si_errno = 0;
info.si_code = SI_KERNEL;
info.si_pid = retval;
/* Note that we don't recalc_sigpending. It is not needed, because
bad things would have happened if CR_SIGNUM was already pending. */
CR_SIGNAL_LOCK(task);
if ((proc_req->saved_sa.sa.sa_handler == SIG_ERR) &&
(proc_req->forced_sa.sa.sa_handler != SIG_ERR)) {
// Force our saved handler (if any), in case user has overwritten.
// Yes, there is a very small race before we actually send the signal.
proc_req->saved_sa = CR_SIGACTION(task, CR_SIGNUM);
CR_SIGACTION(task, CR_SIGNUM) = proc_req->forced_sa;
}
sigdelset(&task->blocked, CR_SIGNUM);
CR_SIGNAL_UNLOCK(task);
retval = send_sig_info(CR_SIGNUM, &info, task);
CR_KTRACE_LOW_LVL("triggered pid %d (%s) w/ retval=%d", task->pid, task->comm, retval);
/* Save the "stoppedness" of the task for later restore */
cr_task->stopped = (task->state == TASK_STOPPED);
if (!retval && (cr_task->stopped)) {
signal_wake_up(task,1);
}
cr_task->self_exec_id = task->self_exec_id;
out:
return retval;
}
//
// __cr_trigger_phase1_only(req)
//
// Sends a trigger for phase1:
// All Phase1 tasks (if any) are sent the trigger.
// The caller must ensure that Phase2 triggers are sent
// Must hold tasklist_lock for reading
//
int __cr_trigger_phase1_only(cr_chkpt_proc_req_t *proc_req)
{
cr_task_t *cr_task;
int found = 0;
list_for_each_entry(cr_task, &proc_req->tasks, proc_req_list) {
if (cr_task->phase == CR_PHASE1) {
found = 1;
do_trigger(cr_task, proc_req);
// XXX: check error conditions
}
}
return found;
}
//
// __cr_trigger_phase1(proc_req)
//
// Sends a trigger for phase1:
// All Phase1 tasks (if any) are sent the trigger.
// If no Phase1 tasks exist then skip to Phase2.
// Must hold req->lock, or otherwise ensure exclusive access to req->tasks
// Must hold tasklist_lock for reading
//
// XXX: Should (probably) never fail, thus currently always returns 0.
//
int __cr_trigger_phase1(cr_chkpt_proc_req_t *proc_req)
{
int found;
found = __cr_trigger_phase1_only(proc_req);
// If no Phase1 tasks were found in this proc, then skip to Phase2
if (!found) {
__cr_trigger_phase2(proc_req);
}
return 0;
}
//
// cr_trigger_phase1(req)
//
// Sends a trigger for phase1:
// All Phase1 tasks (if any) are sent the trigger.
// If no Phase1 tasks exist then skip to Phase2.
//
// XXX: Should (probably) never fail, thus currently always returns 0.
//
int cr_trigger_phase1(cr_chkpt_req_t *req)
{
cr_chkpt_proc_req_t *proc_req;
int retval = 0;
read_lock(&req->lock);
read_lock(&tasklist_lock);
// Send trigger to all of the Phase1 tasks on the list
list_for_each_entry(proc_req, &req->procs, list) {
retval = __cr_trigger_phase1(proc_req);
if (retval) break;
}
read_unlock(&tasklist_lock);
read_unlock(&req->lock);
return retval;
}
//
// __cr_trigger_phase2(proc_req)
//
// Sends the trigger the Phase2 (ie normal/application) tasks.
// Must hold req->lock, or otherwise ensure exclusive access to req->tasks
// Must hold tasklist_lock for reading
//
// XXX: Should (probably) never fail, thus currently always returns 0.
//
int __cr_trigger_phase2(cr_chkpt_proc_req_t *proc_req)
{
cr_task_t *cr_task;
// Send trigger to all of the Phase2 (and "undeclared") tasks.
list_for_each_entry(cr_task, &proc_req->tasks, proc_req_list) {
if (cr_task->phase != CR_PHASE1) {
do_trigger(cr_task, proc_req);
// XXX: check error conditions
}
}
return 0;
}
//
// cr_trigger_phase2(req, proc_req)
//
// Sends the trigger the Phase2 (ie normal/application) tasks.
//
// XXX: Should (probably) never fail, thus currently always returns 0.
//
int cr_trigger_phase2(cr_chkpt_req_t *req, cr_chkpt_proc_req_t *proc_req)
{
int retval;
read_lock(&req->lock);
read_lock(&tasklist_lock);
retval = __cr_trigger_phase2(proc_req);
read_unlock(&tasklist_lock);
read_unlock(&req->lock);
return retval;
}
|