aboutsummaryrefslogtreecommitdiffstats
path: root/arch/s390/appldata/appldata_os.c
blob: 161acc5c8a1b875a05d2ab0d2e97fe38e1dfb1e2 (plain) (blame)
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
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
/*
 * arch/s390/appldata/appldata_os.c
 *
 * Data gathering module for Linux-VM Monitor Stream, Stage 1.
 * Collects misc. OS related data (CPU utilization, running processes).
 *
 * Copyright (C) 2003,2006 IBM Corporation, IBM Deutschland Entwicklung GmbH.
 *
 * Author: Gerald Schaefer <gerald.schaefer@de.ibm.com>
 */

#include <linux/module.h>
#include <linux/init.h>
#include <linux/slab.h>
#include <linux/errno.h>
#include <linux/kernel_stat.h>
#include <linux/netdevice.h>
#include <linux/sched.h>
#include <asm/smp.h>

#include "appldata.h"


#define MY_PRINT_NAME	"appldata_os"		/* for debug messages, etc. */
#define LOAD_INT(x) ((x) >> FSHIFT)
#define LOAD_FRAC(x) LOAD_INT(((x) & (FIXED_1-1)) * 100)

/*
 * OS data
 *
 * This is accessed as binary data by z/VM. If changes to it can't be avoided,
 * the structure version (product ID, see appldata_base.c) needs to be changed
 * as well and all documentation and z/VM applications using it must be
 * updated.
 *
 * The record layout is documented in the Linux for zSeries Device Drivers
 * book:
 * http://oss.software.ibm.com/developerworks/opensource/linux390/index.shtml
 */
struct appldata_os_per_cpu {
	u32 per_cpu_user;	/* timer ticks spent in user mode   */
	u32 per_cpu_nice;	/* ... spent with modified priority */
	u32 per_cpu_system;	/* ... spent in kernel mode         */
	u32 per_cpu_idle;	/* ... spent in idle mode           */

	/* New in 2.6 */
	u32 per_cpu_irq;	/* ... spent in interrupts          */
	u32 per_cpu_softirq;	/* ... spent in softirqs            */
	u32 per_cpu_iowait;	/* ... spent while waiting for I/O  */

	/* New in modification level 01 */
	u32 per_cpu_steal;	/* ... stolen by hypervisor	    */
	u32 cpu_id;		/* number of this CPU		    */
} __attribute__((packed));

struct appldata_os_data {
	u64 timestamp;
	u32 sync_count_1;	/* after VM collected the record data, */
	u32 sync_count_2;	/* sync_count_1 and sync_count_2 should be the
				   same. If not, the record has been updated on
				   the Linux side while VM was collecting the
				   (possibly corrupt) data */

	u32 nr_cpus;		/* number of (virtual) CPUs        */
	u32 per_cpu_size;	/* size of the per-cpu data struct */
	u32 cpu_offset;		/* offset of the first per-cpu data struct */

	u32 nr_running;		/* number of runnable threads      */
	u32 nr_threads;		/* number of threads               */
	u32 avenrun[3];		/* average nr. of running processes during */
				/* the last 1, 5 and 15 minutes */

	/* New in 2.6 */
	u32 nr_iowait;		/* number of blocked threads
				   (waiting for I/O)               */

	/* per cpu data */
	struct appldata_os_per_cpu os_cpu[0];
} __attribute__((packed));

static struct appldata_os_data *appldata_os_data;

static struct appldata_ops ops = {
	.ctl_nr    = CTL_APPLDATA_OS,
	.name	   = "os",
	.record_nr = APPLDATA_RECORD_OS_ID,
	.owner	   = THIS_MODULE,
	.mod_lvl   = {0xF0, 0xF1},		/* EBCDIC "01" */
};


static inline void appldata_print_debug(struct appldata_os_data *os_data)
{
	int a0, a1, a2, i;

	P_DEBUG("--- OS - RECORD ---\n");
	P_DEBUG("nr_threads   = %u\n", os_data->nr_threads);
	P_DEBUG("nr_running   = %u\n", os_data->nr_running);
	P_DEBUG("nr_iowait    = %u\n", os_data->nr_iowait);
	P_DEBUG("avenrun(int) = %8x / %8x / %8x\n", os_data->avenrun[0],
		os_data->avenrun[1], os_data->avenrun[2]);
	a0 = os_data->avenrun[0];
	a1 = os_data->avenrun[1];
	a2 = os_data->avenrun[2];
	P_DEBUG("avenrun(float) = %d.%02d / %d.%02d / %d.%02d\n",
		LOAD_INT(a0), LOAD_FRAC(a0), LOAD_INT(a1), LOAD_FRAC(a1),
		LOAD_INT(a2), LOAD_FRAC(a2));

	P_DEBUG("nr_cpus = %u\n", os_data->nr_cpus);
	for (i = 0; i < os_data->nr_cpus; i++) {
		P_DEBUG("cpu%u : user = %u, nice = %u, system = %u, "
			"idle = %u, irq = %u, softirq = %u, iowait = %u, "
			"steal = %u\n",
				os_data->os_cpu[i].cpu_id,
				os_data->os_cpu[i].per_cpu_user,
				os_data->os_cpu[i].per_cpu_nice,
				os_data->os_cpu[i].per_cpu_system,
				os_data->os_cpu[i].per_cpu_idle,
				os_data->os_cpu[i].per_cpu_irq,
				os_data->os_cpu[i].per_cpu_softirq,
				os_data->os_cpu[i].per_cpu_iowait,
				os_data->os_cpu[i].per_cpu_steal);
	}

	P_DEBUG("sync_count_1 = %u\n", os_data->sync_count_1);
	P_DEBUG("sync_count_2 = %u\n", os_data->sync_count_2);
	P_DEBUG("timestamp    = %lX\n", os_data->timestamp);
}

/*
 * appldata_get_os_data()
 *
 * gather OS data
 */
static void appldata_get_os_data(void *data)
{
	int i, j, rc;
	struct appldata_os_data *os_data;
	unsigned int new_size;

	os_data = data;
	os_data->sync_count_1++;

	os_data->nr_threads = nr_threads;
	os_data->nr_running = nr_running();
	os_data->nr_iowait  = nr_iowait();
	os_data->avenrun[0] = avenrun[0] + (FIXED_1/200);
	os_data->avenrun[1] = avenrun[1] + (FIXED_1/200);
	os_data->avenrun[2] = avenrun[2] + (FIXED_1/200);

	j = 0;
	for_each_online_cpu(i) {
		os_data->os_cpu[j].per_cpu_user =
			cputime_to_jiffies(kstat_cpu(i).cpustat.user);
		os_data->os_cpu[j].per_cpu_nice =
			cputime_to_jiffies(kstat_cpu(i).cpustat.nice);
		os_data->os_cpu[j].per_cpu_system =
			cputime_to_jiffies(kstat_cpu(i).cpustat.system);
		os_data->os_cpu[j].per_cpu_idle =
			cputime_to_jiffies(kstat_cpu(i).cpustat.idle);
		os_data->os_cpu[j].per_cpu_irq =
			cputime_to_jiffies(kstat_cpu(i).cpustat.irq);
		os_data->os_cpu[j].per_cpu_softirq =
			cputime_to_jiffies(kstat_cpu(i).cpustat.softirq);
		os_data->os_cpu[j].per_cpu_iowait =
			cputime_to_jiffies(kstat_cpu(i).cpustat.iowait);
		os_data->os_cpu[j].per_cpu_steal =
			cputime_to_jiffies(kstat_cpu(i).cpustat.steal);
		os_data->os_cpu[j].cpu_id = i;
		j++;
	}

	os_data->nr_cpus = j;

	new_size = sizeof(struct appldata_os_data) +
		   (os_data->nr_cpus * sizeof(struct appldata_os_per_cpu));
	if (ops.size != new_size) {
		if (ops.active) {
			rc = appldata_diag(APPLDATA_RECORD_OS_ID,
					   APPLDATA_START_INTERVAL_REC,
					   (unsigned long) ops.data, new_size,
					   ops.mod_lvl);
			if (rc != 0) {
				P_ERROR("os: START NEW DIAG 0xDC failed, "
					"return code: %d, new size = %i\n", rc,
					new_size);
				P_INFO("os: stopping old record now\n");
			} else
				P_INFO("os: new record size = %i\n", new_size);

			rc = appldata_diag(APPLDATA_RECORD_OS_ID,
					   APPLDATA_STOP_REC,
					   (unsigned long) ops.data, ops.size,
					   ops.mod_lvl);
			if (rc != 0)
				P_ERROR("os: STOP OLD DIAG 0xDC failed, "
					"return code: %d, old size = %i\n", rc,
					ops.size);
			else
				P_INFO("os: old record size = %i stopped\n",
					ops.size);
		}
		ops.size = new_size;
	}
	os_data->timestamp = get_clock();
	os_data->sync_count_2++;
#ifdef APPLDATA_DEBUG
	appldata_print_debug(os_data);
#endif
}


/*
 * appldata_os_init()
 *
 * init data, register ops
 */
static int __init appldata_os_init(void)
{
	int rc, max_size;

	max_size = sizeof(struct appldata_os_data) +
		   (NR_CPUS * sizeof(struct appldata_os_per_cpu));
	if (max_size > APPLDATA_MAX_REC_SIZE) {
		P_ERROR("Max. size of OS record = %i, bigger than maximum "
			"record size (%i)\n", max_size, APPLDATA_MAX_REC_SIZE);
		rc = -ENOMEM;
		goto out;
	}
	P_DEBUG("max. sizeof(os) = %i, sizeof(os_cpu) = %lu\n", max_size,
		sizeof(struct appldata_os_per_cpu));

	appldata_os_data = kzalloc(max_size, GFP_DMA);
	if (appldata_os_data == NULL) {
		P_ERROR("No memory for %s!\n", ops.name);
		rc = -ENOMEM;
		goto out;
	}

	appldata_os_data->per_cpu_size = sizeof(struct appldata_os_per_cpu);
	appldata_os_data->cpu_offset   = offsetof(struct appldata_os_data,
							os_cpu);
	P_DEBUG("cpu offset = %u\n", appldata_os_data->cpu_offset);

	ops.data = appldata_os_data;
	ops.callback  = &appldata_get_os_data;
	rc = appldata_register_ops(&ops);
	if (rc != 0) {
		P_ERROR("Error registering ops, rc = %i\n", rc);
		kfree(appldata_os_data);
	} else {
		P_DEBUG("%s-ops registered!\n", ops.name);
	}
out:
	return rc;
}

/*
 * appldata_os_exit()
 *
 * unregister ops
 */
static void __exit appldata_os_exit(void)
{
	appldata_unregister_ops(&ops);
	kfree(appldata_os_data);
	P_DEBUG("%s-ops unregistered!\n", ops.name);
}


module_init(appldata_os_init);
module_exit(appldata_os_exit);

MODULE_LICENSE("GPL");
MODULE_AUTHOR("Gerald Schaefer");
MODULE_DESCRIPTION("Linux-VM Monitor Stream, OS statistics");