OpenBSD manual page server

Manual Page Search Parameters
KTRACE(2) System Calls Manual KTRACE(2)

ktrace — process tracing

#include <sys/types.h>
#include <sys/param.h>
#include <sys/uio.h>
#include <sys/ktrace.h>

int
ktrace(const char *tracefile, int ops, int trpoints, pid_t pid);

int
fktrace(int tracefd, int ops, int trpoints, pid_t pid);

The () function enables or disables tracing of one or more processes. Users may only trace their own processes. Only the superuser can trace setuid or setgid programs. ktrace() is only available on kernels compiled with the KTRACE option.

tracefile gives the pathname of the file to be used for tracing. The file must exist, be writable by the calling process, and not be a symbolic link. An existing file descriptor tracefd may be used with (). All trace records are always appended to the file, so the file must be truncated to zero length to discard previous trace data. If tracing points are being disabled (see KTROP_CLEAR below), tracefile may be NULL or tracefd may be -1.

The ops parameter specifies the requested ktrace operation. The defined operations are:

Enable trace points specified in trpoints.
Disable trace points specified in trpoints.
Stop all tracing.
The tracing change should apply to the specified process and all its current children.

The trpoints parameter specifies the trace points of interest. The defined trace points are:

Trace system calls.
Trace return values from system calls.
Trace name lookup operations.
Trace all I/O (note that this option can generate much output).
Trace posted signals.
Trace various structs
Trace user data coming from utrace(2) calls.
Trace argument vector in execve(2) calls.
Trace environment vector in execve(2) calls.
Trace violations of pledge(2) restrictions.
Inherit tracing to future children.

The pid parameter refers to a process ID. If it is negative, it refers to a process group ID.

Each tracing event outputs a record composed of a generic header followed by a trace point specific structure. The generic header is:

struct ktr_header {
	uint	ktr_type;		/* trace record type */
	pid_t	ktr_pid;		/* process id */
	pid_t	ktr_tid;		/* thread id */
	struct	timespec ktr_time;	/* timestamp */
	char	ktr_comm[MAXCOMLEN+1];	/* command name */
	size_t	ktr_len;		/* length of buf */
};

The ktr_len field specifies the length of the ktr_type data that follows this header. The ktr_pid, ktr_tid, and ktr_comm fields specify the process, thread, and command generating the record. The ktr_time field gives the time (with nanosecond resolution) that the record was generated.

The generic header is followed by ktr_len bytes of a ktr_type record. The type specific records are defined in the <sys/ktrace.h> include file.

Upon successful completion, the value 0 is returned; otherwise the value -1 is returned and the global variable errno is set to indicate the error.

ktrace() will fail if:

[]
A component of the path prefix is not a directory.
[]
No trace points were selected.
[]
A component of a pathname exceeded NAME_MAX characters, or an entire pathname (including the terminating NUL) exceeded PATH_MAX bytes.
[]
The named tracefile does not exist.
[]
Search permission is denied for a component of the path prefix or the path refers to a symbolic link.
[]
Too many symbolic links were encountered in translating the pathname.
[]
An I/O error occurred while reading from or writing to the file system.
[]
No process can be found corresponding to that specified by pid.

kdump(1), ktrace(1), utrace(2)

A ktrace() function call first appeared in 4.4BSD.

August 13, 2017 OpenBSD-6.2