procps/proc/stat.c

564 lines
16 KiB
C
Raw Normal View History

library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
/*
* libprocps - Library to read proc filesystem
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this library; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
#include <errno.h>
#include <fcntl.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
#include <unistd.h>
#include <sys/stat.h>
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
#include <sys/types.h>
#include <proc/stat.h>
#include "procps-private.h"
#define STAT_FILE "/proc/stat"
struct stat_data {
struct procps_jiffs cpu;
unsigned int intr;
unsigned int ctxt;
unsigned int btime;
unsigned int procs;
unsigned int procs_blocked;
unsigned int procs_running;
};
struct procps_jiffs_private {
struct procps_jiffs_hist cpu;
// future additions go here, please
};
struct procps_stat {
int refcount;
int stat_fd;
struct stat_data data;
int jiff_hists_alloc;
int jiff_hists_inuse;
struct procps_jiffs_private cpu_summary;
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
struct procps_jiffs_private *jiff_hists;
};
/*
* procps_stat_new:
*
* Create a new container to hold the stat information
*
* The initial refcount is 1, and needs to be decremented
* to release the resources of the structure.
*
* Returns: a new stat info container
*/
PROCPS_EXPORT int procps_stat_new (
struct procps_stat **info)
{
struct procps_stat *v;
v = calloc(1, sizeof(struct procps_stat));
if (!v)
return -ENOMEM;
v->refcount = 1;
v->stat_fd = -1;
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
/* v->jiff_hists_alloc = 0; unnecessary with calloc, */
/* v->jiff_hists_inuse = 0; but serves as a reminder */
*info = v;
return 0;
}
/*
* procps_stat_read:
*
* Read the data out of /proc/stat putting the information
* into the supplied info structure.
*
* If CPU stats only needed, set cpu_only to non-zero
*/
PROCPS_EXPORT int procps_stat_read (
struct procps_stat *info,
const int cpu_only)
{
char buf[8192];
char *head, *tail;
int size;
if (info == NULL)
return -EINVAL;
memset(&(info->data), 0, sizeof(struct stat_data));
/* read in the data */
if (-1 == info->stat_fd && (info->stat_fd = open(STAT_FILE, O_RDONLY)) == -1) {
return -errno;
}
if (lseek(info->stat_fd, 0L, SEEK_SET) == -1) {
return -errno;
}
for (;;) {
if ((size = read(info->stat_fd, buf, sizeof(buf)-1)) < 0) {
if (errno == EINTR || errno == EAGAIN)
continue;
return -errno;
}
break;
}
if (size == 0)
return 0;
buf[size] = '\0';
/* Scan the file */
head = buf;
do {
tail = strchr(head, ' ');
if (!tail)
break;
*tail = '\0';
if (0 == strcmp(head, "cpu")) {
if (sscanf(tail+1, "%llu %llu %llu %llu %llu %llu %llu %llu %llu %llu",
&(info->data.cpu.user),
&(info->data.cpu.nice),
&(info->data.cpu.system),
&(info->data.cpu.idle),
&(info->data.cpu.iowait),
&(info->data.cpu.irq),
&(info->data.cpu.sirq),
&(info->data.cpu.stolen),
&(info->data.cpu.guest),
&(info->data.cpu.nice)
) != 10)
return -1;
if (cpu_only)
return 0; // we got what we need
} else if (0 == strcmp(head, "intr")) {
info->data.intr = strtoul(tail+1, &tail, 10);
} else if (0 == strcmp(head, "ctxt")) {
info->data.ctxt = strtoul(tail+1, &tail, 10);
} else if (0 == strcmp(head, "btime")) {
info->data.btime = strtoul(tail+1, &tail, 10);
} else if (0 == strcmp(head, "processes")) {
info->data.procs = strtoul(tail+1, &tail, 10);
} else if (0 == strcmp(head, "procs_blocked")) {
info->data.procs_blocked = strtoul(tail+1, &tail, 10);
} else if (0 == strcmp(head, "procs_running")) {
info->data.procs_running = strtoul(tail+1, &tail, 10);
}
if (tail[0] != '\n')
tail = strchr(tail+1, '\n');
if (!tail)
break;
head = tail + 1;
} while (tail);
if (info->data.procs)
info->data.procs--; // exclude this process
return 0;
}
PROCPS_EXPORT int procps_stat_ref (
struct procps_stat *info)
{
if (info == NULL)
return -EINVAL;
info->refcount++;
return info->refcount;
}
PROCPS_EXPORT int procps_stat_unref (
struct procps_stat **info)
{
if (info == NULL || *info == NULL)
return -EINVAL;
(*info)->refcount--;
if ((*info)->refcount == 0) {
if ((*info)->jiff_hists != NULL)
free((*info)->jiff_hists);
free(*info);
*info = NULL;
return 0;
}
return (*info)->refcount;
}
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
PROCPS_EXPORT jiff procps_stat_cpu_get (
struct procps_stat *info,
enum procps_cpu_item item)
{
switch (item) {
case PROCPS_CPU_USER:
return info->data.cpu.user;
case PROCPS_CPU_NICE:
return info->data.cpu.nice;
case PROCPS_CPU_SYSTEM:
return info->data.cpu.system;
case PROCPS_CPU_IDLE:
return info->data.cpu.idle;
case PROCPS_CPU_IOWAIT:
return info->data.cpu.iowait;
case PROCPS_CPU_IRQ:
return info->data.cpu.irq;
case PROCPS_CPU_SIRQ:
return info->data.cpu.sirq;
case PROCPS_CPU_STOLEN:
return info->data.cpu.stolen;
case PROCPS_CPU_GUEST:
return info->data.cpu.guest;
case PROCPS_CPU_GNICE:
return info->data.cpu.gnice;
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
default:
return 0;
}
}
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
PROCPS_EXPORT int procps_stat_cpu_getstack (
struct procps_stat *info,
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
struct stat_result *these)
{
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
if (these == NULL)
return -EINVAL;
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
for (;;) {
switch (these->item) {
case PROCPS_CPU_USER:
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
these->result.jiff = info->data.cpu.user;
break;
case PROCPS_CPU_NICE:
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
these->result.jiff = info->data.cpu.nice;
break;
case PROCPS_CPU_SYSTEM:
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
these->result.jiff = info->data.cpu.system;
break;
case PROCPS_CPU_IDLE:
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
these->result.jiff = info->data.cpu.idle;
break;
case PROCPS_CPU_IOWAIT:
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
these->result.jiff = info->data.cpu.iowait;
break;
case PROCPS_CPU_IRQ:
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
these->result.jiff = info->data.cpu.irq;
break;
case PROCPS_CPU_SIRQ:
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
these->result.jiff = info->data.cpu.sirq;
break;
case PROCPS_CPU_STOLEN:
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
these->result.jiff = info->data.cpu.stolen;
break;
case PROCPS_CPU_GUEST:
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
these->result.jiff = info->data.cpu.guest;
break;
case PROCPS_CPU_GNICE:
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
these->result.jiff = info->data.cpu.gnice;
break;
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
case PROCPS_CPU_noop:
// don't disturb potential user data in the result struct
break;
case PROCPS_CPU_stack_end:
return 0;
default:
return -EINVAL;
}
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
++these;
}
}
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
PROCPS_EXPORT unsigned int procps_stat_sys_get (
struct procps_stat *info,
enum procps_stat_item item)
{
switch (item) {
case PROCPS_STAT_INTR:
return info->data.intr;
case PROCPS_STAT_CTXT:
return info->data.ctxt;
case PROCPS_STAT_BTIME:
return info->data.btime;
case PROCPS_STAT_PROCS:
return info->data.procs;
case PROCPS_STAT_PROCS_BLK:
return info->data.procs_blocked;
case PROCPS_STAT_PROCS_RUN:
return info->data.procs_running;
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
default:
return 0;
}
}
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
PROCPS_EXPORT int procps_stat_sys_getstack (
struct procps_stat *info,
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
struct stat_result *these)
{
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
if (these == NULL)
return -EINVAL;
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
for (;;) {
switch (these->item) {
case PROCPS_STAT_INTR:
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
these->result.u_int = info->data.intr;
break;
case PROCPS_STAT_CTXT:
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
these->result.u_int = info->data.ctxt;
break;
case PROCPS_STAT_BTIME:
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
these->result.u_int = info->data.btime;
break;
case PROCPS_STAT_PROCS:
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
these->result.u_int = info->data.procs;
break;
case PROCPS_STAT_PROCS_BLK:
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
these->result.u_int = info->data.procs_blocked;
break;
case PROCPS_STAT_PROCS_RUN:
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
these->result.u_int = info->data.procs_running;
break;
case PROCPS_STAT_noop:
// don't disturb potential user data in the result struct
break;
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
case PROCPS_STAT_stack_end:
return 0;
default:
return -EINVAL;
}
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
++these;
}
}
/*
* procps_stat_read_jiffs:
*
* Read the cpu data out of /proc/stat putting the information
* into the dynamically acquired 'jiff_hists' hanging off the
* info structure. Along the way we gather historical stats and
* and embed the cpu summary line in the info structure as well
* ( since we had to read that first /proc/stat line anyway ).
*
* This is all private information but can be copied to caller
* supplied structures upon request.
*/
PROCPS_EXPORT int procps_stat_read_jiffs (
struct procps_stat *info)
{
#define ALLOCincr 32
struct procps_jiffs_private *sum_ptr, *cpu_ptr;
char buf[8192], *bp;
int i, rc, size;
if (info == NULL)
return -EINVAL;
if (!info->jiff_hists_alloc) {
info->jiff_hists = calloc(ALLOCincr, sizeof(struct procps_jiffs_private));
if (!(info->jiff_hists))
return -ENOMEM;
info->jiff_hists_alloc = ALLOCincr;
info->jiff_hists_inuse = 0;
}
if (-1 == info->stat_fd && (info->stat_fd = open(STAT_FILE, O_RDONLY)) == -1)
return -errno;
if (lseek(info->stat_fd, 0L, SEEK_SET) == -1)
return -errno;
for (;;) {
if ((size = read(info->stat_fd, buf, sizeof(buf)-1)) < 0) {
if (errno == EINTR || errno == EAGAIN)
continue;
return -errno;
}
break;
}
if (size == 0)
return 0;
buf[size] = '\0';
bp = buf;
sum_ptr = &info->cpu_summary;
// remember from last time around
memcpy(&sum_ptr->cpu.old, &sum_ptr->cpu.new, sizeof(struct procps_jiffs));
// then value the summary line
if (8 > sscanf(bp, "cpu %llu %llu %llu %llu %llu %llu %llu %llu %llu %llu"
, &sum_ptr->cpu.new.user, &sum_ptr->cpu.new.nice, &sum_ptr->cpu.new.system
, &sum_ptr->cpu.new.idle, &sum_ptr->cpu.new.iowait, &sum_ptr->cpu.new.irq
, &sum_ptr->cpu.new.sirq, &sum_ptr->cpu.new.stolen
, &sum_ptr->cpu.new.guest, &sum_ptr->cpu.new.gnice))
return -1;
sum_ptr->cpu.id = -1; // mark as summary
i = 0;
reap_em_again:
cpu_ptr = info->jiff_hists + i; // adapt to relocated if reap_em_again
do {
bp = 1 + strchr(bp, '\n');
// remember from last time around
memcpy(&cpu_ptr->cpu.old, &cpu_ptr->cpu.new, sizeof(struct procps_jiffs));
if (8 > (rc = sscanf(bp, "cpu%d %llu %llu %llu %llu %llu %llu %llu %llu %llu %llu"
, &cpu_ptr->cpu.id
, &cpu_ptr->cpu.new.user, &cpu_ptr->cpu.new.nice, &cpu_ptr->cpu.new.system
, &cpu_ptr->cpu.new.idle, &cpu_ptr->cpu.new.iowait, &cpu_ptr->cpu.new.irq
, &cpu_ptr->cpu.new.sirq, &cpu_ptr->cpu.new.stolen
, &cpu_ptr->cpu.new.guest, &cpu_ptr->cpu.new.gnice))) {
memmove(cpu_ptr, sum_ptr, sizeof(struct procps_jiffs_hist));
break; // we must tolerate cpus taken offline
}
++i;
++cpu_ptr;
} while (i < info->jiff_hists_alloc);
if (i == info->jiff_hists_alloc && rc >= 8) {
info->jiff_hists_alloc += ALLOCincr;
info->jiff_hists = realloc(info->jiff_hists, info->jiff_hists_alloc * sizeof(struct procps_jiffs_private));
if (!(info->jiff_hists))
return -ENOMEM;
goto reap_em_again;
}
info->jiff_hists_inuse = i;
return i;
#undef ALLOCincr
}
/*
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
* procps_stat_jiffs_get:
*
* Return the designated cpu data in the caller supplied structure.
* A negative 'which' denotes the cpu_summary, not a real cpu.
*
* This function deals only with the 'current' jiffs counts.
*/
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
PROCPS_EXPORT int procps_stat_jiffs_get (
struct procps_stat *info,
struct procps_jiffs *dest,
int which)
{
struct procps_jiffs_private *p;
int i;
if (info == NULL || dest == NULL)
return -EINVAL;
if (which < 0) {
// note, we're just copying the 'new' portion of our procps_jiffs_private
memcpy(dest, &info->cpu_summary, sizeof(struct procps_jiffs));
return 0;
}
p = info->jiff_hists;
for (i = 0; i < info->jiff_hists_inuse; i++) {
if (p->cpu.id == which) {
// note, we're just copying the 'new' portion of our procps_jiffs_private
memcpy(dest, p, sizeof(struct procps_jiffs));
return 0;
}
++p;
}
return -1;
}
/*
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
* procps_stat_jiffs_fill:
*
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
* Refresh available cpu data, then return all cpu data in the caller
* supplied structures, up to the lesser of maxdests or total available.
*
* We tolerate a maxdests greater than the total available, and
* the caller had better tolerate fewer returned than requested.
*
* This function deals only with the 'current' jiffs counts.
*/
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
PROCPS_EXPORT int procps_stat_jiffs_fill (
struct procps_stat *info,
struct procps_jiffs *dests,
int maxdests)
{
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
int i, rc;
if (info == NULL || dests == NULL)
return -EINVAL;
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
if ((rc = procps_stat_read_jiffs(info)) < 0)
return rc;
if (!info->jiff_hists_inuse)
return -1;
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
for (i = 0; i < info->jiff_hists_inuse && i < maxdests; i++) {
// note, we're just copying the 'new' portion of our procps_jiffs_private
memcpy(dests + i, info->jiff_hists + i, sizeof(struct procps_jiffs));
}
return i;
}
/*
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
* procps_stat_jiffs_hist_get:
*
* Return the designated cpu data in the caller supplied structure.
* A negative 'which' denotes the cpu_summary, not a real cpu.
*
* This function provides both 'new' and 'old' jiffs counts.
*/
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
PROCPS_EXPORT int procps_stat_jiffs_hist_get (
struct procps_stat *info,
struct procps_jiffs_hist *dest,
int which)
{
struct procps_jiffs_private *p;
int i;
if (info == NULL || dest == NULL)
return -EINVAL;
if (which < 0) {
memcpy(dest, &info->cpu_summary, sizeof(struct procps_jiffs_hist));
return 0;
}
p = info->jiff_hists;
for (i = 0; i < info->jiff_hists_inuse; i++) {
if (p->cpu.id == which) {
memcpy(dest, p, sizeof(struct procps_jiffs_hist));
return 0;
}
++p;
}
return -1;
}
/*
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
* procps_stat_jiffs_hist_fill:
*
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
* Refresh available cpu data, then return all cpu data in the caller
* supplied structures, up to the lesser of maxdests or total available.
*
* We tolerate a maxdests greater than the total available, and
* the caller had better tolerate fewer returned than requested.
*
* This function provides both 'new' and 'old' jiffs counts.
*/
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
PROCPS_EXPORT int procps_stat_jiffs_hist_fill (
struct procps_stat *info,
struct procps_jiffs_hist *dests,
int maxdests)
{
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
int i, rc;
if (info == NULL || dests == NULL)
return -EINVAL;
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
if ((rc = procps_stat_read_jiffs(info)) < 0)
return rc;
if (!info->jiff_hists_inuse)
return -1;
library: readstat redesigned using 'stack' vs. 'chain' In addition to that text shown below the line which is common to several commit messages, this patch contains several minor changes with lessor impact upon the API: . A call to procps_stat_read_jiffs() has been added to those jiffs functions carrying the 'fill' nomenclature to parallel like functions in some of our other files. . The #include header files are ordered alphabetically now, with all those <sys/??> types separately grouped. . Standard copyright boilerplate was added in .c file. . The header file follows the conventions of indenting (by 4 spaces) those parameters too lengthy for 1 line. ------------------------------------------------------ . The former 'chains' have now become 'stacks' without the 'next' pointer in each result struct. The pointers initially seemed to offer some flexibility with memory allocations and benefits for the library access logic. However, user access was always via displacement and a a statically allocated chain was cumbersome to define. . An enumerator ending in '_noop' will no longer serve as a fencepost delimiter. Rather, it has become a much more important and flexible user oriented tool. Adding one or more such 'items' in any items list passed into the library becomes the means of extending the 'stack' to also include user (not just library) data. Any such data is guaranteed to never be altered by the library. . Anticipating PID support, where many different types must be represented in a result structure, we'll adopt a common naming standard. And, while not every results structure currently needs to reflect disparate types a union will be employed so the same dot qualifier ('.') can be used consistently when accessing all such data. Signed-off-by: Jim Warner <james.warner@comcast.net>
2015-07-21 10:30:00 +05:30
for (i = 0; i < info->jiff_hists_inuse && i < maxdests; i++) {
memcpy(dests + i, info->jiff_hists + i, sizeof(struct procps_jiffs_hist));
}
return i;
}