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
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
|
// SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
// Copyright (c) 2022, Huawei
#include "vmlinux.h"
#include <bpf/bpf_helpers.h>
#include <bpf/bpf_tracing.h>
#include <bpf/bpf_core_read.h>
/*
* This should be in sync with "util/kwork.h"
*/
enum kwork_class_type {
KWORK_CLASS_IRQ,
KWORK_CLASS_SOFTIRQ,
KWORK_CLASS_WORKQUEUE,
KWORK_CLASS_SCHED,
KWORK_CLASS_MAX,
};
#define MAX_ENTRIES 102400
#define MAX_NR_CPUS 2048
#define PF_KTHREAD 0x00200000
#define MAX_COMMAND_LEN 16
struct time_data {
__u64 timestamp;
};
struct work_data {
__u64 runtime;
};
struct task_data {
__u32 tgid;
__u32 is_kthread;
char comm[MAX_COMMAND_LEN];
};
struct work_key {
__u32 type;
__u32 pid;
__u64 task_p;
};
struct task_key {
__u32 pid;
__u32 cpu;
};
struct {
__uint(type, BPF_MAP_TYPE_TASK_STORAGE);
__uint(map_flags, BPF_F_NO_PREALLOC);
__type(key, int);
__type(value, struct time_data);
} kwork_top_task_time SEC(".maps");
struct {
__uint(type, BPF_MAP_TYPE_PERCPU_HASH);
__uint(key_size, sizeof(struct work_key));
__uint(value_size, sizeof(struct time_data));
__uint(max_entries, MAX_ENTRIES);
} kwork_top_irq_time SEC(".maps");
struct {
__uint(type, BPF_MAP_TYPE_HASH);
__uint(key_size, sizeof(struct task_key));
__uint(value_size, sizeof(struct task_data));
__uint(max_entries, MAX_ENTRIES);
} kwork_top_tasks SEC(".maps");
struct {
__uint(type, BPF_MAP_TYPE_PERCPU_HASH);
__uint(key_size, sizeof(struct work_key));
__uint(value_size, sizeof(struct work_data));
__uint(max_entries, MAX_ENTRIES);
} kwork_top_works SEC(".maps");
struct {
__uint(type, BPF_MAP_TYPE_HASH);
__uint(key_size, sizeof(u32));
__uint(value_size, sizeof(u8));
__uint(max_entries, MAX_NR_CPUS);
} kwork_top_cpu_filter SEC(".maps");
int enabled = 0;
int has_cpu_filter = 0;
__u64 from_timestamp = 0;
__u64 to_timestamp = 0;
static __always_inline int cpu_is_filtered(__u32 cpu)
{
__u8 *cpu_val;
if (has_cpu_filter) {
cpu_val = bpf_map_lookup_elem(&kwork_top_cpu_filter, &cpu);
if (!cpu_val)
return 1;
}
return 0;
}
static __always_inline void update_task_info(struct task_struct *task, __u32 cpu)
{
struct task_key key = {
.pid = task->pid,
.cpu = cpu,
};
if (!bpf_map_lookup_elem(&kwork_top_tasks, &key)) {
struct task_data data = {
.tgid = task->tgid,
.is_kthread = task->flags & PF_KTHREAD ? 1 : 0,
};
BPF_CORE_READ_STR_INTO(&data.comm, task, comm);
bpf_map_update_elem(&kwork_top_tasks, &key, &data, BPF_ANY);
}
}
static __always_inline void update_work(struct work_key *key, __u64 delta)
{
struct work_data *data;
data = bpf_map_lookup_elem(&kwork_top_works, key);
if (data) {
data->runtime += delta;
} else {
struct work_data new_data = {
.runtime = delta,
};
bpf_map_update_elem(&kwork_top_works, key, &new_data, BPF_ANY);
}
}
static void on_sched_out(struct task_struct *task, __u64 ts, __u32 cpu)
{
__u64 delta;
struct time_data *pelem;
pelem = bpf_task_storage_get(&kwork_top_task_time, task, NULL, 0);
if (pelem)
delta = ts - pelem->timestamp;
else
delta = ts - from_timestamp;
struct work_key key = {
.type = KWORK_CLASS_SCHED,
.pid = task->pid,
.task_p = (__u64)task,
};
update_work(&key, delta);
update_task_info(task, cpu);
}
static void on_sched_in(struct task_struct *task, __u64 ts)
{
struct time_data *pelem;
pelem = bpf_task_storage_get(&kwork_top_task_time, task, NULL,
BPF_LOCAL_STORAGE_GET_F_CREATE);
if (pelem)
pelem->timestamp = ts;
}
SEC("tp_btf/sched_switch")
int on_switch(u64 *ctx)
{
struct task_struct *prev, *next;
prev = (struct task_struct *)ctx[1];
next = (struct task_struct *)ctx[2];
if (!enabled)
return 0;
__u32 cpu = bpf_get_smp_processor_id();
if (cpu_is_filtered(cpu))
return 0;
__u64 ts = bpf_ktime_get_ns();
on_sched_out(prev, ts, cpu);
on_sched_in(next, ts);
return 0;
}
SEC("tp_btf/irq_handler_entry")
int on_irq_handler_entry(u64 *cxt)
{
struct task_struct *task;
if (!enabled)
return 0;
__u32 cpu = bpf_get_smp_processor_id();
if (cpu_is_filtered(cpu))
return 0;
__u64 ts = bpf_ktime_get_ns();
task = (struct task_struct *)bpf_get_current_task();
if (!task)
return 0;
struct work_key key = {
.type = KWORK_CLASS_IRQ,
.pid = BPF_CORE_READ(task, pid),
.task_p = (__u64)task,
};
struct time_data data = {
.timestamp = ts,
};
bpf_map_update_elem(&kwork_top_irq_time, &key, &data, BPF_ANY);
return 0;
}
SEC("tp_btf/irq_handler_exit")
int on_irq_handler_exit(u64 *cxt)
{
__u64 delta;
struct task_struct *task;
struct time_data *pelem;
if (!enabled)
return 0;
__u32 cpu = bpf_get_smp_processor_id();
if (cpu_is_filtered(cpu))
return 0;
__u64 ts = bpf_ktime_get_ns();
task = (struct task_struct *)bpf_get_current_task();
if (!task)
return 0;
struct work_key key = {
.type = KWORK_CLASS_IRQ,
.pid = BPF_CORE_READ(task, pid),
.task_p = (__u64)task,
};
pelem = bpf_map_lookup_elem(&kwork_top_irq_time, &key);
if (pelem && pelem->timestamp != 0)
delta = ts - pelem->timestamp;
else
delta = ts - from_timestamp;
update_work(&key, delta);
return 0;
}
SEC("tp_btf/softirq_entry")
int on_softirq_entry(u64 *cxt)
{
struct task_struct *task;
if (!enabled)
return 0;
__u32 cpu = bpf_get_smp_processor_id();
if (cpu_is_filtered(cpu))
return 0;
__u64 ts = bpf_ktime_get_ns();
task = (struct task_struct *)bpf_get_current_task();
if (!task)
return 0;
struct work_key key = {
.type = KWORK_CLASS_SOFTIRQ,
.pid = BPF_CORE_READ(task, pid),
.task_p = (__u64)task,
};
struct time_data data = {
.timestamp = ts,
};
bpf_map_update_elem(&kwork_top_irq_time, &key, &data, BPF_ANY);
return 0;
}
SEC("tp_btf/softirq_exit")
int on_softirq_exit(u64 *cxt)
{
__u64 delta;
struct task_struct *task;
struct time_data *pelem;
if (!enabled)
return 0;
__u32 cpu = bpf_get_smp_processor_id();
if (cpu_is_filtered(cpu))
return 0;
__u64 ts = bpf_ktime_get_ns();
task = (struct task_struct *)bpf_get_current_task();
if (!task)
return 0;
struct work_key key = {
.type = KWORK_CLASS_SOFTIRQ,
.pid = BPF_CORE_READ(task, pid),
.task_p = (__u64)task,
};
pelem = bpf_map_lookup_elem(&kwork_top_irq_time, &key);
if (pelem)
delta = ts - pelem->timestamp;
else
delta = ts - from_timestamp;
update_work(&key, delta);
return 0;
}
char LICENSE[] SEC("license") = "Dual BSD/GPL";
|
