#include <linux/sched.h>
#include <linux/pid_namespace.h>
#include <linux/binfmts.h>
-#include <linux/version.h>
-#if (LINUX_VERSION_CODE >= KERNEL_VERSION(3,9,0))
#include <linux/sched/rt.h>
-#endif
#define LTTNG_MAX_PID_NS_LEVEL 32
#ifndef _TRACE_SCHED_DEF_
#define _TRACE_SCHED_DEF_
-#if (LINUX_VERSION_CODE >= KERNEL_VERSION(4,15,0))
-
static inline long __trace_sched_switch_state(bool preempt, struct task_struct *p)
{
unsigned int state;
return state ? (1 << (state - 1)) : state;
}
-#elif (LINUX_VERSION_CODE >= KERNEL_VERSION(4,14,0))
-
-static inline long __trace_sched_switch_state(bool preempt, struct task_struct *p)
-{
- unsigned int state;
-
-#ifdef CONFIG_SCHED_DEBUG
- BUG_ON(p != current);
-#endif /* CONFIG_SCHED_DEBUG */
-
- /*
- * Preemption ignores task state, therefore preempted tasks are always
- * RUNNING (we will not have dequeued if state != RUNNING).
- */
- if (preempt)
- return TASK_REPORT_MAX;
-
- /*
- * __get_task_state() uses fls() and returns a value from 0-8 range.
- * Decrement it by 1 (except TASK_RUNNING state i.e 0) before using
- * it for left shift operation to get the correct task->state
- * mapping.
- */
- state = __get_task_state(p);
-
- return state ? (1 << (state - 1)) : state;
-}
-
-#elif (LINUX_VERSION_CODE >= KERNEL_VERSION(4,4,0))
-
-static inline long __trace_sched_switch_state(bool preempt, struct task_struct *p)
-{
-#ifdef CONFIG_SCHED_DEBUG
- BUG_ON(p != current);
-#endif /* CONFIG_SCHED_DEBUG */
- /*
- * Preemption ignores task state, therefore preempted tasks are always RUNNING
- * (we will not have dequeued if state != RUNNING).
- */
- return preempt ? TASK_RUNNING | TASK_STATE_MAX : p->state;
-}
-
-#elif (LINUX_VERSION_CODE >= KERNEL_VERSION(3,19,0))
-
-static inline long __trace_sched_switch_state(struct task_struct *p)
-{
- long state = p->state;
-
-#ifdef CONFIG_PREEMPT
-#ifdef CONFIG_SCHED_DEBUG
- BUG_ON(p != current);
-#endif /* CONFIG_SCHED_DEBUG */
- /*
- * For all intents and purposes a preempted task is a running task.
- */
- if (preempt_count() & PREEMPT_ACTIVE)
- state = TASK_RUNNING | TASK_STATE_MAX;
-#endif /* CONFIG_PREEMPT */
-
- return state;
-}
-
-#elif (LINUX_VERSION_CODE >= KERNEL_VERSION(3,13,0))
-
-static inline long __trace_sched_switch_state(struct task_struct *p)
-{
- long state = p->state;
-
-#ifdef CONFIG_PREEMPT
- /*
- * For all intents and purposes a preempted task is a running task.
- */
- if (task_preempt_count(p) & PREEMPT_ACTIVE)
- state = TASK_RUNNING | TASK_STATE_MAX;
-#endif
-
- return state;
-}
-
-#elif (LINUX_VERSION_CODE >= KERNEL_VERSION(3,2,0))
-
-static inline long __trace_sched_switch_state(struct task_struct *p)
-{
- long state = p->state;
-
-#ifdef CONFIG_PREEMPT
- /*
- * For all intents and purposes a preempted task is a running task.
- */
- if (task_thread_info(p)->preempt_count & PREEMPT_ACTIVE)
- state = TASK_RUNNING | TASK_STATE_MAX;
-#endif
-
- return state;
-}
-
-#else
-
-static inline long __trace_sched_switch_state(struct task_struct *p)
-{
- long state = p->state;
-
-#ifdef CONFIG_PREEMPT
- /*
- * For all intents and purposes a preempted task is a running task.
- */
- if (task_thread_info(p)->preempt_count & PREEMPT_ACTIVE)
- state = TASK_RUNNING;
-#endif
-
- return state;
-}
-
-#endif
-
#endif /* _TRACE_SCHED_DEF_ */
/*
ctf_enum_value("TASK_TRACED", __TASK_TRACED)
ctf_enum_value("EXIT_DEAD", EXIT_DEAD)
ctf_enum_value("EXIT_ZOMBIE", EXIT_ZOMBIE)
-
-#if (LINUX_VERSION_CODE >= KERNEL_VERSION(3,9,0))
ctf_enum_value("TASK_PARKED", TASK_PARKED)
-#endif /* #if (LINUX_VERSION_CODE >= KERNEL_VERSION(3,9,0)) */
-
ctf_enum_value("TASK_DEAD", TASK_DEAD)
ctf_enum_value("TASK_WAKEKILL", TASK_WAKEKILL)
ctf_enum_value("TASK_WAKING", TASK_WAKING)
-
-#if (LINUX_VERSION_CODE >= KERNEL_VERSION(4,2,0))
ctf_enum_value("TASK_NOLOAD", TASK_NOLOAD)
-#endif /* #if (LINUX_VERSION_CODE >= KERNEL_VERSION(4,2,0)) */
-
-#if (LINUX_VERSION_CODE >= KERNEL_VERSION(4,8,0))
ctf_enum_value("TASK_NEW", TASK_NEW)
-#endif /* #if (LINUX_VERSION_CODE >= KERNEL_VERSION(4,8,0)) */
-
ctf_enum_value("TASK_STATE_MAX", TASK_STATE_MAX)
)
)
/*
* Tracepoint for waking up a task:
*/
-#if (LINUX_VERSION_CODE >= KERNEL_VERSION(4,3,0) || \
- LTTNG_RT_KERNEL_RANGE(4,1,10,11, 4,2,0,0) || \
- LTTNG_RT_KERNEL_RANGE(3,18,27,26, 3,19,0,0) || \
- LTTNG_RT_KERNEL_RANGE(3,14,61,63, 3,15,0,0) || \
- LTTNG_RT_KERNEL_RANGE(3,12,54,73, 3,13,0,0) || \
- LTTNG_RT_KERNEL_RANGE(3,10,97,106, 3,11,0,0) || \
- LTTNG_RT_KERNEL_RANGE(3,4,110,139, 3,5,0,0) || \
- LTTNG_RT_KERNEL_RANGE(3,2,77,111, 3,3,0,0))
LTTNG_TRACEPOINT_EVENT_CLASS(sched_wakeup_template,
TP_PROTO(struct task_struct *p),
ctf_integer(int, target_cpu, task_cpu(p))
)
)
-#else /* #if (LINUX_VERSION_CODE >= KERNEL_VERSION(4,3,0)) */
-LTTNG_TRACEPOINT_EVENT_CLASS(sched_wakeup_template,
-
- TP_PROTO(struct task_struct *p, int success),
-
- TP_ARGS(p, success),
-
- TP_FIELDS(
- ctf_array_text(char, comm, p->comm, TASK_COMM_LEN)
- ctf_integer(pid_t, tid, p->pid)
- ctf_integer(int, prio, p->prio - MAX_RT_PRIO)
- ctf_integer(int, success, success)
- ctf_integer(int, target_cpu, task_cpu(p))
- )
-)
-#endif /* #else #if (LINUX_VERSION_CODE >= KERNEL_VERSION(4,3,0)) */
-
-#if (LINUX_VERSION_CODE >= KERNEL_VERSION(4,3,0) || \
- LTTNG_RT_KERNEL_RANGE(4,1,10,11, 4,2,0,0) || \
- LTTNG_RT_KERNEL_RANGE(3,18,27,26, 3,19,0,0) || \
- LTTNG_RT_KERNEL_RANGE(3,14,61,63, 3,15,0,0) || \
- LTTNG_RT_KERNEL_RANGE(3,12,54,73, 3,13,0,0) || \
- LTTNG_RT_KERNEL_RANGE(3,10,97,106, 3,11,0,0) || \
- LTTNG_RT_KERNEL_RANGE(3,4,110,139, 3,5,0,0) || \
- LTTNG_RT_KERNEL_RANGE(3,2,77,111, 3,3,0,0))
/*
* Tracepoint called when waking a task; this tracepoint is guaranteed to be
TP_PROTO(struct task_struct *p),
TP_ARGS(p))
-#else
-
-LTTNG_TRACEPOINT_EVENT_INSTANCE(sched_wakeup_template, sched_wakeup,
- TP_PROTO(struct task_struct *p, int success),
- TP_ARGS(p, success))
-
-/*
- * Tracepoint for waking up a new task:
- */
-LTTNG_TRACEPOINT_EVENT_INSTANCE(sched_wakeup_template, sched_wakeup_new,
- TP_PROTO(struct task_struct *p, int success),
- TP_ARGS(p, success))
-
-#endif /* #if (LINUX_VERSION_CODE >= KERNEL_VERSION(4,3,0)) */
-
/*
* Tracepoint for task switches, performed by the scheduler:
*/
LTTNG_TRACEPOINT_EVENT(sched_switch,
-#if (LINUX_VERSION_CODE >= KERNEL_VERSION(4,4,0))
TP_PROTO(bool preempt,
struct task_struct *prev,
struct task_struct *next),
TP_ARGS(preempt, prev, next),
-#else
- TP_PROTO(struct task_struct *prev,
- struct task_struct *next),
-
- TP_ARGS(prev, next),
-#endif /* #if (LINUX_VERSION_CODE >= KERNEL_VERSION(4,4,0)) */
TP_FIELDS(
ctf_array_text(char, prev_comm, prev->comm, TASK_COMM_LEN)
ctf_integer(pid_t, prev_tid, prev->pid)
ctf_integer(int, prev_prio, prev->prio - MAX_RT_PRIO)
-#if (LINUX_VERSION_CODE >= KERNEL_VERSION(4,4,0))
ctf_enum(task_state, long, prev_state, __trace_sched_switch_state(preempt, prev))
-#else
- ctf_enum(task_state, long, prev_state, __trace_sched_switch_state(prev))
-#endif
ctf_array_text(char, next_comm, next->comm, TASK_COMM_LEN)
ctf_integer(pid_t, next_tid, next->pid)
ctf_integer(int, next_prio, next->prio - MAX_RT_PRIO)
ctf_array_text(char, parent_comm, parent->comm, TASK_COMM_LEN)
ctf_integer(pid_t, parent_tid, parent->pid)
ctf_integer(pid_t, parent_pid, parent->tgid)
-#if (LINUX_VERSION_CODE >= KERNEL_VERSION(3,8,0))
ctf_integer(unsigned int, parent_ns_inum,
({
unsigned int parent_ns_inum = 0;
}
parent_ns_inum;
}))
-#endif
ctf_array_text(char, child_comm, child->comm, TASK_COMM_LEN)
ctf_integer(pid_t, child_tid, child->pid)
ctf_sequence(pid_t, vtids, tp_locvar->vtids, u8, tp_locvar->ns_level)
ctf_integer(pid_t, child_pid, child->tgid)
-#if (LINUX_VERSION_CODE >= KERNEL_VERSION(3,8,0))
ctf_integer(unsigned int, child_ns_inum,
({
unsigned int child_ns_inum = 0;
}
child_ns_inum;
}))
-#endif
),
TP_code_post()
)
-#if (LINUX_VERSION_CODE >= KERNEL_VERSION(3,4,0))
/*
* Tracepoint for exec:
*/
ctf_integer(pid_t, old_tid, old_pid)
)
)
-#endif
/*
* XXX the below sched_stat tracepoints only apply to SCHED_OTHER/BATCH/IDLE
TP_PROTO(struct task_struct *tsk, u64 delay),
TP_ARGS(tsk, delay))
-#if (LINUX_VERSION_CODE >= KERNEL_VERSION(3,3,0))
/*
* Tracepoint for accounting blocked time (time the task is in uninterruptible).
*/
LTTNG_TRACEPOINT_EVENT_INSTANCE(sched_stat_template, sched_stat_blocked,
TP_PROTO(struct task_struct *tsk, u64 delay),
TP_ARGS(tsk, delay))
-#endif
/*
* Tracepoint for accounting runtime (time the task is executing
)
)
-#if (LINUX_VERSION_CODE >= KERNEL_VERSION(4,12,0) || \
- LTTNG_RT_KERNEL_RANGE(4,9,27,18, 4,10,0,0) || \
- LTTNG_RT_KERNEL_RANGE(4,11,5,1, 4,12,0,0))
/*
* Tracepoint for showing priority inheritance modifying a tasks
* priority.
ctf_integer(int, newprio, pi_task ? pi_task->prio - MAX_RT_PRIO : tsk->prio - MAX_RT_PRIO)
)
)
-#else
-/*
- * Tracepoint for showing priority inheritance modifying a tasks
- * priority.
- */
-LTTNG_TRACEPOINT_EVENT(sched_pi_setprio,
-
- TP_PROTO(struct task_struct *tsk, int newprio),
-
- TP_ARGS(tsk, newprio),
-
- TP_FIELDS(
- ctf_array_text(char, comm, tsk->comm, TASK_COMM_LEN)
- ctf_integer(pid_t, tid, tsk->pid)
- ctf_integer(int, oldprio, tsk->prio - MAX_RT_PRIO)
- ctf_integer(int, newprio, newprio - MAX_RT_PRIO)
- )
-)
-#endif
#endif /* LTTNG_TRACE_SCHED_H */