Pfault —
enable and disable the tracing of faults
Process Control Library (libproc, -lproc)
#include
<libproc.h>
int
Pfault(
struct
ps_prochandle *P,
int which,
int stop);
The
Pfault() function controls what the
process
P should do on faults.
A fault is a hardware event that occurs in the context of a running process and
thread. A hardware fault may occur because an illegal instruction was
executed, a breakpoint or watchpoint was encountered, or an arithmetic
exception occurred, among others. The full list of faults is available in both
proc(4) and
<sys/fault.h>.
For each hardware fault, a process may be configured to stop the thread that
encountered it when it occurs. The value of the
stop parameter controls whether or not the
listed fault in
which will cause the thread
to trap. A value of 1 indicates the thread should stop; a value of 0 indicates
it should not.
The value of
which indicates which hardware
fault the change applies to. However, if the value of
which is zero, then it applies to all faults.
The
Pfault() function only applies to
actively running processes. It does not function on handles that refer to core
files, zombie processes, or ELF objects.
Upon successful completion, the
Pfault()
function returns the old disposition of the fault --
0 if it was not set to stop and
1 if it was -- and the fault state is updated.
Otherwise,
-1 is returned,
errno is updated with the error that
occurred, and the fault state is not updated.
The
Pfault() function will fail if:
-
-
EINVAL- The value of which is invalid, e.g. it is
less than zero or greater than the largest defined fault.
-
-
ENOENT- The handle P refers to a process that is
a zombie, a core file, or a file.
Uncommitted
See
LOCKING in
libproc(3LIB).
libproc(3LIB),
proc(4)