Copyright (c) 2005, Sun Microsystems, Inc.
All Rights Reserved.
The contents of this file are subject to the terms of the Common Development and Distribution License (the "License"). You may not use this file except in compliance with the License.
You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE or http://www.opensolaris.org/os/licensing. See the License for the specific language governing permissions and limitations under the License.
When distributing Covered Code, include this CDDL HEADER in each file and include the License file at usr/src/OPENSOLARIS.LICENSE. If applicable, add the following below this CDDL HEADER, with the fields enclosed by brackets "[]" replaced with your own identifying information: Portions Copyright [yyyy] [name of copyright owner]
cc [ flag... ] file... -lcpc [ library... ] #include <libcpc.h> int cpc_count_usr_events(int enable);
int cpc_count_sys_events(int enable);
In certain applications, it can be useful to explicitly enable and disable performance counters at different times so that the performance of a critical algorithm can be examined. The cpc_count_usr_events() function can be used to control whether events are counted on behalf of the application running in user mode, while cpc_count_sys_events() can be used to control whether events are counted on behalf of the application while it is running in the kernel, without otherwise disturbing the binding of events to the invoking LWP. If the enable argument is non-zero, counting of events is enabled, otherwise they are disabled.
Upon successful completion, cpc_count_usr_events() and cpc_count_sys_events() return 0. Otherwise, the functions return -1 and set errno to indicate the error.
The cpc_count_usr_events() and cpc_count_sys_events() functions will fail if: EAGAIN
The associated performance counter context has been invalidated by another process.
No performance counter context has been created, or an attempt was made to enable system events while delivering counter overflow signals.
Example 1 Use cpc_count_usr_events() to minimize code needed by application.
In this example, the routine cpc_count_usr_events() is used to minimize the amount of code that needs to be added to the application. The cputrack(1) command can be used in conjunction with these interfaces to provide event programming, sampling, and reporting facilities.
If the application is instrumented in this way and then started by cputrack with the nouser flag set in the event specification, counting of user events will only be enabled around the critical code section of interest. If the program is run normally, no harm will ensue.
int have_counters = 0;
int
main(int argc, char *argv[])
{
if (cpc_version(CPC_VER_CURRENT) == CPC_VER_CURRENT &&
cpc_getcpuver() != -1 && cpc_access() == 0)
have_counters = 1;
/* ... other application code */
if (have_counters)
(void) cpc_count_usr_events(1);
/* ==> Code to be measured goes here <== */
if (have_counters)
(void) cpc_count_usr_events(0);
/* ... other application code */
}
See attributes(5) for descriptions of the following attributes:
| ATTRIBUTE TYPE ATTRIBUTE VALUE |
| MT-Level MT-Safe |
| Interface Stability Obsolete |
cputrack(1), cpc(3CPC), cpc_access(3CPC), cpc_bind_event(3CPC), cpc_enable(3CPC), cpc_getcpuver(3CPC), cpc_pctx_bind_event(3CPC), cpc_version(3CPC), libcpc(3LIB), attributes(5)
The cpc_count_usr_events() and cpc_count_sys_events() functions exist for binary compatibility only. Source containing these functions will not compile. These functions are obsolete and might be removed in a future release. Applications should use cpc_enable(3CPC) instead.