2015-07-07 18:12:06 +05:30
|
|
|
/*
|
|
|
|
* diskstat - Disk statistics - part of procps
|
|
|
|
*
|
|
|
|
* Copyright (C) 2003 Fabian Frederick
|
|
|
|
* Copyright (C) 2003 Albert Cahalan
|
|
|
|
* Copyright (C) 2015 Craig Small <csmall@enc.com.au>
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
* Copyright (C) 2016 Jim Warner <james.warner@comcast.net>
|
2015-07-07 18:12:06 +05:30
|
|
|
*
|
|
|
|
* 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: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
#include <errno.h>
|
|
|
|
#include <fcntl.h>
|
|
|
|
#include <dirent.h>
|
2015-07-07 18:12:06 +05:30
|
|
|
#include <stdio.h>
|
|
|
|
#include <stdlib.h>
|
|
|
|
#include <string.h>
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
#include <time.h>
|
2015-07-07 18:12:06 +05:30
|
|
|
#include <unistd.h>
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
|
2015-07-07 18:12:06 +05:30
|
|
|
#include <sys/stat.h>
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
#include <sys/types.h>
|
2015-07-07 18:12:06 +05:30
|
|
|
|
|
|
|
#include "procps-private.h"
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
#include <proc/diskstats.h>
|
2015-07-07 18:12:06 +05:30
|
|
|
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
/* The following define will cause the 'node_add' function to maintain our |
|
|
|
|
nodes list in ascending alphabetical order which could be used to avoid |
|
|
|
|
a sort on name. Without it, we default to a 'pull-up' stack at slightly |
|
|
|
|
more effort than a simple 'push-down' list to duplicate prior behavior. | */
|
|
|
|
//#define ALPHABETIC_NODES
|
2015-07-07 18:12:06 +05:30
|
|
|
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
#define DISKSTATS_LINE_LEN 1024
|
|
|
|
#define DISKSTATS_NAME_LEN 34
|
|
|
|
#define DISKSTATS_FILE "/proc/diskstats"
|
|
|
|
#define SYSBLOCK_DIR "/sys/block"
|
|
|
|
|
|
|
|
#define STACKS_INCR 64
|
|
|
|
#define STR_COMPARE strverscmp
|
|
|
|
|
|
|
|
struct dev_data {
|
2015-07-07 18:12:06 +05:30
|
|
|
unsigned long reads;
|
|
|
|
unsigned long reads_merged;
|
|
|
|
unsigned long read_sectors;
|
|
|
|
unsigned long read_time;
|
|
|
|
unsigned long writes;
|
|
|
|
unsigned long writes_merged;
|
|
|
|
unsigned long write_sectors;
|
|
|
|
unsigned long write_time;
|
|
|
|
unsigned long io_inprogress;
|
|
|
|
unsigned long io_time;
|
|
|
|
unsigned long io_wtime;
|
|
|
|
};
|
|
|
|
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
struct dev_node {
|
|
|
|
char name[DISKSTATS_NAME_LEN+1];
|
|
|
|
int type;
|
|
|
|
int major;
|
|
|
|
int minor;
|
|
|
|
time_t stamped;
|
|
|
|
struct dev_data new;
|
|
|
|
struct dev_data old;
|
|
|
|
struct dev_node *next;
|
|
|
|
};
|
|
|
|
|
|
|
|
struct stacks_extent {
|
|
|
|
int ext_numstacks;
|
|
|
|
struct stacks_extent *next;
|
|
|
|
struct diskstats_stack **stacks;
|
|
|
|
};
|
|
|
|
|
|
|
|
struct ext_support {
|
|
|
|
int numitems; // includes 'logical_end' delimiter
|
|
|
|
enum diskstats_item *items; // includes 'logical_end' delimiter
|
|
|
|
struct stacks_extent *extents; // anchor for these extents
|
|
|
|
int dirty_stacks;
|
|
|
|
};
|
|
|
|
|
|
|
|
struct fetch_support {
|
|
|
|
struct diskstats_stack **anchor; // fetch consolidated extents
|
|
|
|
int n_alloc; // number of above pointers allocated
|
|
|
|
int n_inuse; // number of above pointers occupied
|
|
|
|
int n_alloc_save; // last known reap.stacks allocation
|
|
|
|
struct diskstats_reap results; // count + stacks for return to caller
|
|
|
|
};
|
|
|
|
|
2016-07-21 10:30:00 +05:30
|
|
|
struct diskstats_info {
|
2015-07-07 18:12:06 +05:30
|
|
|
int refcount;
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
FILE *diskstats_fp;
|
|
|
|
time_t old_stamp; /* previous read seconds */
|
|
|
|
time_t new_stamp; /* current read seconds */
|
|
|
|
struct dev_node *nodes; /* dev nodes anchor */
|
|
|
|
struct ext_support select_ext; /* supports concurrent select/reap */
|
|
|
|
struct ext_support fetch_ext; /* supports concurrent select/reap */
|
|
|
|
struct fetch_support fetch; /* support for procps_diskstats_reap */
|
|
|
|
struct diskstats_result get_this; /* used by procps_diskstats_get */
|
2015-07-07 18:12:06 +05:30
|
|
|
};
|
|
|
|
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
|
|
|
|
// ___ Results 'Set' Support ||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|
|
|
|
|
|
#define setNAME(e) set_results_ ## e
|
|
|
|
#define setDECL(e) static void setNAME(e) \
|
|
|
|
(struct diskstats_result *R, struct dev_node *N)
|
|
|
|
|
|
|
|
// regular assignment
|
|
|
|
#define DEV_set(e,t,x) setDECL(e) { R->result. t = N-> x; }
|
|
|
|
#define REG_set(e,t,x) setDECL(e) { R->result. t = N->new . x; }
|
|
|
|
// delta assignment
|
|
|
|
#define HST_set(e,t,x) setDECL(e) { R->result. t = ( N->new . x - N->old. x ); }
|
|
|
|
|
|
|
|
setDECL(noop) { (void)R; (void)N; }
|
|
|
|
setDECL(extra) { (void)R; (void)N; }
|
|
|
|
|
|
|
|
DEV_set(DISKSTATS_NAME, str, name)
|
|
|
|
DEV_set(DISKSTATS_TYPE, s_int, type)
|
|
|
|
DEV_set(DISKSTATS_MAJOR, s_int, major)
|
|
|
|
DEV_set(DISKSTATS_MINOR, s_int, minor)
|
|
|
|
|
|
|
|
REG_set(DISKSTATS_READS, ul_int, reads)
|
|
|
|
REG_set(DISKSTATS_READS_MERGED, ul_int, reads_merged)
|
|
|
|
REG_set(DISKSTATS_READ_SECTORS, ul_int, read_sectors)
|
|
|
|
REG_set(DISKSTATS_READ_TIME, ul_int, read_time)
|
|
|
|
REG_set(DISKSTATS_WRITES, ul_int, writes)
|
|
|
|
REG_set(DISKSTATS_WRITES_MERGED, ul_int, writes_merged)
|
|
|
|
REG_set(DISKSTATS_WRITE_SECTORS, ul_int, write_sectors)
|
|
|
|
REG_set(DISKSTATS_WRITE_TIME, ul_int, write_time)
|
|
|
|
REG_set(DISKSTATS_IO_TIME, ul_int, io_time)
|
|
|
|
REG_set(DISKSTATS_IO_WTIME, ul_int, io_wtime)
|
|
|
|
|
|
|
|
REG_set(DISKSTATS_IO_INPROGRESS, s_int, io_inprogress)
|
|
|
|
|
|
|
|
HST_set(DISKSTATS_DELTA_READS, s_int, reads)
|
|
|
|
HST_set(DISKSTATS_DELTA_READS_MERGED, s_int, reads_merged)
|
|
|
|
HST_set(DISKSTATS_DELTA_READ_SECTORS, s_int, read_sectors)
|
|
|
|
HST_set(DISKSTATS_DELTA_READ_TIME, s_int, read_time)
|
|
|
|
HST_set(DISKSTATS_DELTA_WRITES, s_int, writes)
|
|
|
|
HST_set(DISKSTATS_DELTA_WRITES_MERGED, s_int, writes_merged)
|
|
|
|
HST_set(DISKSTATS_DELTA_WRITE_SECTORS, s_int, write_sectors)
|
|
|
|
HST_set(DISKSTATS_DELTA_WRITE_TIME, s_int, write_time)
|
|
|
|
HST_set(DISKSTATS_DELTA_IO_TIME, s_int, io_time)
|
|
|
|
HST_set(DISKSTATS_DELTA_IO_WTIME, s_int, io_wtime)
|
|
|
|
|
|
|
|
#undef setDECL
|
|
|
|
#undef DEV_set
|
|
|
|
#undef REG_set
|
|
|
|
#undef HST_set
|
|
|
|
|
|
|
|
|
|
|
|
// ___ Sorting Support ||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|
|
|
|
|
|
struct sort_parms {
|
|
|
|
int offset;
|
|
|
|
enum diskstats_sort_order order;
|
|
|
|
};
|
|
|
|
|
|
|
|
#define srtNAME(t) sort_results_ ## t
|
|
|
|
#define srtDECL(t) static int srtNAME(t) \
|
|
|
|
(const struct diskstats_stack **A, const struct diskstats_stack **B, struct sort_parms *P)
|
|
|
|
|
|
|
|
srtDECL(s_int) {
|
|
|
|
const struct diskstats_result *a = (*A)->head + P->offset; \
|
|
|
|
const struct diskstats_result *b = (*B)->head + P->offset; \
|
|
|
|
return P->order * (a->result.s_int - b->result.s_int);
|
|
|
|
}
|
|
|
|
|
|
|
|
srtDECL(ul_int) {
|
|
|
|
const struct diskstats_result *a = (*A)->head + P->offset; \
|
|
|
|
const struct diskstats_result *b = (*B)->head + P->offset; \
|
|
|
|
if ( a->result.ul_int > b->result.ul_int ) return P->order > 0 ? 1 : -1; \
|
|
|
|
if ( a->result.ul_int < b->result.ul_int ) return P->order > 0 ? -1 : 1; \
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
srtDECL(str) {
|
|
|
|
const struct diskstats_result *a = (*A)->head + P->offset;
|
|
|
|
const struct diskstats_result *b = (*B)->head + P->offset;
|
|
|
|
return P->order * STR_COMPARE(a->result.str, b->result.str);
|
|
|
|
}
|
|
|
|
|
|
|
|
srtDECL(noop) { \
|
|
|
|
(void)A; (void)B; (void)P; \
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
#undef srtDECL
|
|
|
|
|
|
|
|
|
|
|
|
// ___ Controlling Table ||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|
|
|
|
|
|
typedef void (*SET_t)(struct diskstats_result *, struct dev_node *);
|
|
|
|
#define RS(e) (SET_t)setNAME(e)
|
|
|
|
|
|
|
|
typedef int (*QSR_t)(const void *, const void *, void *);
|
|
|
|
#define QS(t) (QSR_t)srtNAME(t)
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Need it be said?
|
|
|
|
* This table must be kept in the exact same order as
|
|
|
|
* those *enum diskstats_item* guys ! */
|
|
|
|
static struct {
|
|
|
|
SET_t setsfunc; // the actual result setting routine
|
|
|
|
QSR_t sortfunc; // sort cmp func for a specific type
|
|
|
|
} Item_table[] = {
|
|
|
|
/* setsfunc sortfunc
|
|
|
|
---------------------------------- --------- */
|
|
|
|
{ RS(noop), QS(ul_int) },
|
|
|
|
{ RS(extra), QS(noop) },
|
|
|
|
|
|
|
|
{ RS(DISKSTATS_NAME), QS(str) },
|
|
|
|
{ RS(DISKSTATS_TYPE), QS(s_int) },
|
|
|
|
{ RS(DISKSTATS_MAJOR), QS(s_int) },
|
|
|
|
{ RS(DISKSTATS_MINOR), QS(s_int) },
|
|
|
|
|
|
|
|
{ RS(DISKSTATS_READS), QS(ul_int) },
|
|
|
|
{ RS(DISKSTATS_READS_MERGED), QS(ul_int) },
|
|
|
|
{ RS(DISKSTATS_READ_SECTORS), QS(ul_int) },
|
|
|
|
{ RS(DISKSTATS_READ_TIME), QS(ul_int) },
|
|
|
|
{ RS(DISKSTATS_WRITES), QS(ul_int) },
|
|
|
|
{ RS(DISKSTATS_WRITES_MERGED), QS(ul_int) },
|
|
|
|
{ RS(DISKSTATS_WRITE_SECTORS), QS(ul_int) },
|
|
|
|
{ RS(DISKSTATS_WRITE_TIME), QS(ul_int) },
|
|
|
|
{ RS(DISKSTATS_IO_TIME), QS(ul_int) },
|
|
|
|
{ RS(DISKSTATS_IO_WTIME), QS(ul_int) },
|
|
|
|
|
|
|
|
{ RS(DISKSTATS_IO_INPROGRESS), QS(s_int) },
|
|
|
|
|
|
|
|
{ RS(DISKSTATS_DELTA_READS), QS(s_int) },
|
|
|
|
{ RS(DISKSTATS_DELTA_READS_MERGED), QS(s_int) },
|
|
|
|
{ RS(DISKSTATS_DELTA_READ_SECTORS), QS(s_int) },
|
|
|
|
{ RS(DISKSTATS_DELTA_READ_TIME), QS(s_int) },
|
|
|
|
{ RS(DISKSTATS_DELTA_WRITES), QS(s_int) },
|
|
|
|
{ RS(DISKSTATS_DELTA_WRITES_MERGED), QS(s_int) },
|
|
|
|
{ RS(DISKSTATS_DELTA_WRITE_SECTORS), QS(s_int) },
|
|
|
|
{ RS(DISKSTATS_DELTA_WRITE_TIME), QS(s_int) },
|
|
|
|
{ RS(DISKSTATS_DELTA_IO_TIME), QS(s_int) },
|
|
|
|
{ RS(DISKSTATS_DELTA_IO_WTIME), QS(s_int) },
|
|
|
|
|
library: removed all the 'PROCPS_' enumerator prefixes
Many of our item enumerator identifiers are very long,
especially in that <VMSTAT> module. Additionally, they
all contain the exact same universal 'PROCPS_' prefix.
The origins for this are likely found in the desire to
avoid name clashes with other potential include files.
But with procps-ng newlib, we've probably gone way too
far. Did 'PROCPS_PIDS_TICS_SYSTEM' actually offer more
protection against clash than 'PIDS_TICS_SYSTEM' does?
I don't think so. Besides, no matter how big that name
becomes, one can never guarantee they'll never be some
clash. And, conversely, extremely short names will not
always create conflict. Of course, in either case when
some clash occurs, one can always #undef that problem.
Thus, this commit will eliminate that 'PROCPS_' prefix
making all of those enum identifiers a little shorter.
And, we'll still be well above some ridiculously short
(criminally short) names found in some common headers:
- - - - - - - - - - <term.h>
- 'tab', 'TTY', etc
- - - - - - - - - - - - - - - - <search.h>
- 'ENTER', ENTRY', 'FIND', etc
------------------------------------------------------
Finally, with this as a last of the wholesale changes,
we will have established the naming conventions below:
. only functions will begin with that 'procps_' prefix
. exposed structures begin with the module/header name
. item enumerators begin like structs, but capitalized
. other enumerators work exactly like item enumerators
. macros and constants begin just like the enumerators
------------------------------------------------------
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-21 10:30:00 +05:30
|
|
|
// dummy entry corresponding to DISKSTATS_logical_end ...
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
{ NULL, NULL }
|
|
|
|
};
|
|
|
|
|
|
|
|
/* please note,
|
|
|
|
* this enum MUST be 1 greater than the highest value of any enum */
|
library: removed all the 'PROCPS_' enumerator prefixes
Many of our item enumerator identifiers are very long,
especially in that <VMSTAT> module. Additionally, they
all contain the exact same universal 'PROCPS_' prefix.
The origins for this are likely found in the desire to
avoid name clashes with other potential include files.
But with procps-ng newlib, we've probably gone way too
far. Did 'PROCPS_PIDS_TICS_SYSTEM' actually offer more
protection against clash than 'PIDS_TICS_SYSTEM' does?
I don't think so. Besides, no matter how big that name
becomes, one can never guarantee they'll never be some
clash. And, conversely, extremely short names will not
always create conflict. Of course, in either case when
some clash occurs, one can always #undef that problem.
Thus, this commit will eliminate that 'PROCPS_' prefix
making all of those enum identifiers a little shorter.
And, we'll still be well above some ridiculously short
(criminally short) names found in some common headers:
- - - - - - - - - - <term.h>
- 'tab', 'TTY', etc
- - - - - - - - - - - - - - - - <search.h>
- 'ENTER', ENTRY', 'FIND', etc
------------------------------------------------------
Finally, with this as a last of the wholesale changes,
we will have established the naming conventions below:
. only functions will begin with that 'procps_' prefix
. exposed structures begin with the module/header name
. item enumerators begin like structs, but capitalized
. other enumerators work exactly like item enumerators
. macros and constants begin just like the enumerators
------------------------------------------------------
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-21 10:30:00 +05:30
|
|
|
enum diskstats_item DISKSTATS_logical_end = DISKSTATS_DELTA_IO_WTIME + 1;
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
|
|
|
|
#undef setNAME
|
|
|
|
#undef srtNAME
|
|
|
|
#undef RS
|
|
|
|
#undef QS
|
|
|
|
|
|
|
|
|
|
|
|
// ___ Private Functions ||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|
|
// --- dev_node specific support ----------------------------------------------
|
|
|
|
|
|
|
|
static struct dev_node *node_add (
|
2016-07-21 10:30:00 +05:30
|
|
|
struct diskstats_info *info,
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
struct dev_node *this)
|
|
|
|
{
|
|
|
|
struct dev_node *prev, *walk;
|
|
|
|
|
|
|
|
#ifdef ALPHABETIC_NODES
|
|
|
|
if (!info->nodes
|
|
|
|
|| (STR_COMPARE(this->name, info->nodes->name) < 0)) {
|
|
|
|
this->next = info->nodes;
|
|
|
|
info->nodes = this;
|
|
|
|
return this;
|
|
|
|
}
|
|
|
|
prev = info->nodes;
|
|
|
|
walk = info->nodes->next;
|
|
|
|
while (walk) {
|
|
|
|
if (STR_COMPARE(this->name, walk->name) < 0)
|
|
|
|
break;
|
|
|
|
prev = walk;
|
|
|
|
walk = walk->next;
|
|
|
|
}
|
|
|
|
prev->next = this;
|
|
|
|
this->next = walk;
|
|
|
|
#else
|
|
|
|
if (!info->nodes)
|
|
|
|
info->nodes = this;
|
|
|
|
else {
|
|
|
|
walk = info->nodes;
|
|
|
|
do {
|
|
|
|
prev = walk;
|
|
|
|
walk = walk->next;
|
|
|
|
} while (walk);
|
|
|
|
prev->next = this;
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
return this;
|
|
|
|
} // end: node_add
|
|
|
|
|
2016-07-24 10:30:00 +05:30
|
|
|
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
static void node_classify (
|
|
|
|
struct dev_node *this)
|
2015-07-07 18:12:06 +05:30
|
|
|
{
|
|
|
|
DIR *dirp;
|
|
|
|
struct dirent *dent;
|
|
|
|
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
/* all disks start off as partitions. this function
|
|
|
|
checks /sys/block and changes a device found there
|
|
|
|
into a disk. if /sys/block cannot have the directory
|
|
|
|
read, all devices are then treated as disks. */
|
library: removed all the 'PROCPS_' enumerator prefixes
Many of our item enumerator identifiers are very long,
especially in that <VMSTAT> module. Additionally, they
all contain the exact same universal 'PROCPS_' prefix.
The origins for this are likely found in the desire to
avoid name clashes with other potential include files.
But with procps-ng newlib, we've probably gone way too
far. Did 'PROCPS_PIDS_TICS_SYSTEM' actually offer more
protection against clash than 'PIDS_TICS_SYSTEM' does?
I don't think so. Besides, no matter how big that name
becomes, one can never guarantee they'll never be some
clash. And, conversely, extremely short names will not
always create conflict. Of course, in either case when
some clash occurs, one can always #undef that problem.
Thus, this commit will eliminate that 'PROCPS_' prefix
making all of those enum identifiers a little shorter.
And, we'll still be well above some ridiculously short
(criminally short) names found in some common headers:
- - - - - - - - - - <term.h>
- 'tab', 'TTY', etc
- - - - - - - - - - - - - - - - <search.h>
- 'ENTER', ENTRY', 'FIND', etc
------------------------------------------------------
Finally, with this as a last of the wholesale changes,
we will have established the naming conventions below:
. only functions will begin with that 'procps_' prefix
. exposed structures begin with the module/header name
. item enumerators begin like structs, but capitalized
. other enumerators work exactly like item enumerators
. macros and constants begin just like the enumerators
------------------------------------------------------
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-21 10:30:00 +05:30
|
|
|
this->type = DISKSTATS_TYPE_PARTITION;
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
|
|
|
|
if (!(dirp = opendir(SYSBLOCK_DIR))) {
|
library: removed all the 'PROCPS_' enumerator prefixes
Many of our item enumerator identifiers are very long,
especially in that <VMSTAT> module. Additionally, they
all contain the exact same universal 'PROCPS_' prefix.
The origins for this are likely found in the desire to
avoid name clashes with other potential include files.
But with procps-ng newlib, we've probably gone way too
far. Did 'PROCPS_PIDS_TICS_SYSTEM' actually offer more
protection against clash than 'PIDS_TICS_SYSTEM' does?
I don't think so. Besides, no matter how big that name
becomes, one can never guarantee they'll never be some
clash. And, conversely, extremely short names will not
always create conflict. Of course, in either case when
some clash occurs, one can always #undef that problem.
Thus, this commit will eliminate that 'PROCPS_' prefix
making all of those enum identifiers a little shorter.
And, we'll still be well above some ridiculously short
(criminally short) names found in some common headers:
- - - - - - - - - - <term.h>
- 'tab', 'TTY', etc
- - - - - - - - - - - - - - - - <search.h>
- 'ENTER', ENTRY', 'FIND', etc
------------------------------------------------------
Finally, with this as a last of the wholesale changes,
we will have established the naming conventions below:
. only functions will begin with that 'procps_' prefix
. exposed structures begin with the module/header name
. item enumerators begin like structs, but capitalized
. other enumerators work exactly like item enumerators
. macros and constants begin just like the enumerators
------------------------------------------------------
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-21 10:30:00 +05:30
|
|
|
this->type = DISKSTATS_TYPE_DISK;
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
return;
|
|
|
|
}
|
2016-07-24 10:30:00 +05:30
|
|
|
while ((dent = readdir(dirp))) {
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
if (strcmp(this->name, dent->d_name) == 0) {
|
library: removed all the 'PROCPS_' enumerator prefixes
Many of our item enumerator identifiers are very long,
especially in that <VMSTAT> module. Additionally, they
all contain the exact same universal 'PROCPS_' prefix.
The origins for this are likely found in the desire to
avoid name clashes with other potential include files.
But with procps-ng newlib, we've probably gone way too
far. Did 'PROCPS_PIDS_TICS_SYSTEM' actually offer more
protection against clash than 'PIDS_TICS_SYSTEM' does?
I don't think so. Besides, no matter how big that name
becomes, one can never guarantee they'll never be some
clash. And, conversely, extremely short names will not
always create conflict. Of course, in either case when
some clash occurs, one can always #undef that problem.
Thus, this commit will eliminate that 'PROCPS_' prefix
making all of those enum identifiers a little shorter.
And, we'll still be well above some ridiculously short
(criminally short) names found in some common headers:
- - - - - - - - - - <term.h>
- 'tab', 'TTY', etc
- - - - - - - - - - - - - - - - <search.h>
- 'ENTER', ENTRY', 'FIND', etc
------------------------------------------------------
Finally, with this as a last of the wholesale changes,
we will have established the naming conventions below:
. only functions will begin with that 'procps_' prefix
. exposed structures begin with the module/header name
. item enumerators begin like structs, but capitalized
. other enumerators work exactly like item enumerators
. macros and constants begin just like the enumerators
------------------------------------------------------
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-21 10:30:00 +05:30
|
|
|
this->type = DISKSTATS_TYPE_DISK;
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
break;
|
|
|
|
}
|
2015-07-07 18:12:06 +05:30
|
|
|
}
|
2016-05-03 17:28:13 +05:30
|
|
|
closedir(dirp);
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
} // end: node_classify
|
|
|
|
|
2015-07-07 18:12:06 +05:30
|
|
|
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
static struct dev_node *node_cut (
|
2016-07-21 10:30:00 +05:30
|
|
|
struct diskstats_info *info,
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
struct dev_node *this)
|
2015-07-07 18:12:06 +05:30
|
|
|
{
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
struct dev_node *node = info->nodes;
|
|
|
|
|
|
|
|
if (this) {
|
|
|
|
if (this == node) {
|
|
|
|
info->nodes = node->next;
|
|
|
|
return this;
|
|
|
|
}
|
|
|
|
do {
|
|
|
|
if (this == node->next) {
|
|
|
|
node->next = node->next->next;
|
|
|
|
return this;
|
|
|
|
}
|
|
|
|
node = node->next;
|
|
|
|
} while (node);
|
2015-07-07 18:12:06 +05:30
|
|
|
}
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
return NULL;
|
|
|
|
} // end: node_cut
|
|
|
|
|
2015-07-07 18:12:06 +05:30
|
|
|
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
static struct dev_node *node_get (
|
2016-07-21 10:30:00 +05:30
|
|
|
struct diskstats_info *info,
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
const char *name)
|
2015-07-07 18:12:06 +05:30
|
|
|
{
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
struct dev_node *node = info->nodes;
|
|
|
|
|
|
|
|
while (node) {
|
|
|
|
if (strcmp(name, node->name) == 0)
|
|
|
|
break;
|
|
|
|
node = node->next;
|
|
|
|
}
|
|
|
|
if (node) {
|
|
|
|
/* if this disk or partition has somehow gotten stale, we'll lose
|
|
|
|
it and then pretend it was never actually found ...
|
|
|
|
[ we test against both stamps in case a 'read' was avoided ] */
|
|
|
|
if (node->stamped != info->old_stamp
|
|
|
|
&& (node->stamped != info->new_stamp)) {
|
|
|
|
free(node_cut(info, node));
|
|
|
|
node = NULL;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return node;
|
|
|
|
} // end: node_get
|
2015-07-07 18:12:06 +05:30
|
|
|
|
|
|
|
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
static int node_update (
|
2016-07-21 10:30:00 +05:30
|
|
|
struct diskstats_info *info,
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
struct dev_node *source)
|
|
|
|
{
|
|
|
|
struct dev_node *target = node_get(info, source->name);
|
|
|
|
|
|
|
|
if (!target) {
|
|
|
|
if (!(target = malloc(sizeof(struct dev_node))))
|
|
|
|
return -ENOMEM;
|
|
|
|
memcpy(target, source, sizeof(struct dev_node));
|
|
|
|
// let's not distort the deltas when a new node is created ...
|
|
|
|
memcpy(&target->old, &target->new, sizeof(struct dev_data));
|
|
|
|
node_classify(target);
|
|
|
|
node_add(info, target);
|
2015-07-07 18:12:06 +05:30
|
|
|
return 0;
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
}
|
|
|
|
// remember history from last time around ...
|
|
|
|
memcpy(&source->old, &target->new, sizeof(struct dev_data));
|
|
|
|
// preserve some stuff from the existing node struct ...
|
|
|
|
source->type = target->type;
|
|
|
|
source->next = target->next;
|
|
|
|
// finally 'update' the existing node struct ...
|
|
|
|
memcpy(target, source, sizeof(struct dev_node));
|
|
|
|
return 0;
|
|
|
|
} // end: node_update
|
2015-07-07 18:12:06 +05:30
|
|
|
|
|
|
|
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
// ___ Private Functions ||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|
|
// --- generalized support ----------------------------------------------------
|
2015-07-07 18:12:06 +05:30
|
|
|
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
static inline void assign_results (
|
|
|
|
struct diskstats_stack *stack,
|
|
|
|
struct dev_node *node)
|
2015-07-07 18:12:06 +05:30
|
|
|
{
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
struct diskstats_result *this = stack->head;
|
|
|
|
|
|
|
|
for (;;) {
|
|
|
|
enum diskstats_item item = this->item;
|
library: removed all the 'PROCPS_' enumerator prefixes
Many of our item enumerator identifiers are very long,
especially in that <VMSTAT> module. Additionally, they
all contain the exact same universal 'PROCPS_' prefix.
The origins for this are likely found in the desire to
avoid name clashes with other potential include files.
But with procps-ng newlib, we've probably gone way too
far. Did 'PROCPS_PIDS_TICS_SYSTEM' actually offer more
protection against clash than 'PIDS_TICS_SYSTEM' does?
I don't think so. Besides, no matter how big that name
becomes, one can never guarantee they'll never be some
clash. And, conversely, extremely short names will not
always create conflict. Of course, in either case when
some clash occurs, one can always #undef that problem.
Thus, this commit will eliminate that 'PROCPS_' prefix
making all of those enum identifiers a little shorter.
And, we'll still be well above some ridiculously short
(criminally short) names found in some common headers:
- - - - - - - - - - <term.h>
- 'tab', 'TTY', etc
- - - - - - - - - - - - - - - - <search.h>
- 'ENTER', ENTRY', 'FIND', etc
------------------------------------------------------
Finally, with this as a last of the wholesale changes,
we will have established the naming conventions below:
. only functions will begin with that 'procps_' prefix
. exposed structures begin with the module/header name
. item enumerators begin like structs, but capitalized
. other enumerators work exactly like item enumerators
. macros and constants begin just like the enumerators
------------------------------------------------------
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-21 10:30:00 +05:30
|
|
|
if (item >= DISKSTATS_logical_end)
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
break;
|
|
|
|
Item_table[item].setsfunc(this, node);
|
|
|
|
++this;
|
|
|
|
}
|
|
|
|
return;
|
|
|
|
} // end: assign_results
|
2015-07-07 18:12:06 +05:30
|
|
|
|
|
|
|
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
static inline void cleanup_stack (
|
|
|
|
struct diskstats_result *this)
|
|
|
|
{
|
|
|
|
for (;;) {
|
library: removed all the 'PROCPS_' enumerator prefixes
Many of our item enumerator identifiers are very long,
especially in that <VMSTAT> module. Additionally, they
all contain the exact same universal 'PROCPS_' prefix.
The origins for this are likely found in the desire to
avoid name clashes with other potential include files.
But with procps-ng newlib, we've probably gone way too
far. Did 'PROCPS_PIDS_TICS_SYSTEM' actually offer more
protection against clash than 'PIDS_TICS_SYSTEM' does?
I don't think so. Besides, no matter how big that name
becomes, one can never guarantee they'll never be some
clash. And, conversely, extremely short names will not
always create conflict. Of course, in either case when
some clash occurs, one can always #undef that problem.
Thus, this commit will eliminate that 'PROCPS_' prefix
making all of those enum identifiers a little shorter.
And, we'll still be well above some ridiculously short
(criminally short) names found in some common headers:
- - - - - - - - - - <term.h>
- 'tab', 'TTY', etc
- - - - - - - - - - - - - - - - <search.h>
- 'ENTER', ENTRY', 'FIND', etc
------------------------------------------------------
Finally, with this as a last of the wholesale changes,
we will have established the naming conventions below:
. only functions will begin with that 'procps_' prefix
. exposed structures begin with the module/header name
. item enumerators begin like structs, but capitalized
. other enumerators work exactly like item enumerators
. macros and constants begin just like the enumerators
------------------------------------------------------
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-21 10:30:00 +05:30
|
|
|
if (this->item >= DISKSTATS_logical_end)
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
break;
|
library: removed all the 'PROCPS_' enumerator prefixes
Many of our item enumerator identifiers are very long,
especially in that <VMSTAT> module. Additionally, they
all contain the exact same universal 'PROCPS_' prefix.
The origins for this are likely found in the desire to
avoid name clashes with other potential include files.
But with procps-ng newlib, we've probably gone way too
far. Did 'PROCPS_PIDS_TICS_SYSTEM' actually offer more
protection against clash than 'PIDS_TICS_SYSTEM' does?
I don't think so. Besides, no matter how big that name
becomes, one can never guarantee they'll never be some
clash. And, conversely, extremely short names will not
always create conflict. Of course, in either case when
some clash occurs, one can always #undef that problem.
Thus, this commit will eliminate that 'PROCPS_' prefix
making all of those enum identifiers a little shorter.
And, we'll still be well above some ridiculously short
(criminally short) names found in some common headers:
- - - - - - - - - - <term.h>
- 'tab', 'TTY', etc
- - - - - - - - - - - - - - - - <search.h>
- 'ENTER', ENTRY', 'FIND', etc
------------------------------------------------------
Finally, with this as a last of the wholesale changes,
we will have established the naming conventions below:
. only functions will begin with that 'procps_' prefix
. exposed structures begin with the module/header name
. item enumerators begin like structs, but capitalized
. other enumerators work exactly like item enumerators
. macros and constants begin just like the enumerators
------------------------------------------------------
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-21 10:30:00 +05:30
|
|
|
if (this->item > DISKSTATS_noop)
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
this->result.ul_int = 0;
|
|
|
|
++this;
|
2015-07-07 18:12:06 +05:30
|
|
|
}
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
} // end: cleanup_stack
|
2015-07-07 18:12:06 +05:30
|
|
|
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
|
|
|
|
static inline void cleanup_stacks_all (
|
|
|
|
struct ext_support *this)
|
2015-07-07 18:12:06 +05:30
|
|
|
{
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
struct stacks_extent *ext = this->extents;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
while (ext) {
|
|
|
|
for (i = 0; ext->stacks[i]; i++)
|
|
|
|
cleanup_stack(ext->stacks[i]->head);
|
|
|
|
ext = ext->next;
|
|
|
|
};
|
|
|
|
this->dirty_stacks = 0;
|
|
|
|
} // end: cleanup_stacks_all
|
|
|
|
|
|
|
|
|
|
|
|
static void extents_free_all (
|
|
|
|
struct ext_support *this)
|
|
|
|
{
|
|
|
|
while (this->extents) {
|
|
|
|
struct stacks_extent *p = this->extents;
|
|
|
|
this->extents = this->extents->next;
|
|
|
|
free(p);
|
|
|
|
};
|
|
|
|
} // end: extents_free_all
|
|
|
|
|
|
|
|
|
|
|
|
static inline struct diskstats_result *itemize_stack (
|
|
|
|
struct diskstats_result *p,
|
|
|
|
int depth,
|
|
|
|
enum diskstats_item *items)
|
|
|
|
{
|
|
|
|
struct diskstats_result *p_sav = p;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
for (i = 0; i < depth; i++) {
|
|
|
|
p->item = items[i];
|
|
|
|
p->result.ul_int = 0;
|
|
|
|
++p;
|
|
|
|
}
|
|
|
|
return p_sav;
|
|
|
|
} // end: itemize_stack
|
|
|
|
|
|
|
|
|
|
|
|
static void itemize_stacks_all (
|
|
|
|
struct ext_support *this)
|
|
|
|
{
|
|
|
|
struct stacks_extent *ext = this->extents;
|
|
|
|
|
|
|
|
while (ext) {
|
|
|
|
int i;
|
|
|
|
for (i = 0; ext->stacks[i]; i++)
|
|
|
|
itemize_stack(ext->stacks[i]->head, this->numitems, this->items);
|
|
|
|
ext = ext->next;
|
|
|
|
};
|
|
|
|
this->dirty_stacks = 0;
|
|
|
|
} // end: static void itemize_stacks_all
|
|
|
|
|
|
|
|
|
|
|
|
static inline int items_check_failed (
|
|
|
|
struct ext_support *this,
|
|
|
|
enum diskstats_item *items,
|
|
|
|
int numitems)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
|
|
|
/* if an enum is passed instead of an address of one or more enums, ol' gcc
|
|
|
|
* will silently convert it to an address (possibly NULL). only clang will
|
|
|
|
* offer any sort of warning like the following:
|
|
|
|
*
|
|
|
|
* warning: incompatible integer to pointer conversion passing 'int' to parameter of type 'enum diskstats_item *'
|
library: removed all the 'PROCPS_' enumerator prefixes
Many of our item enumerator identifiers are very long,
especially in that <VMSTAT> module. Additionally, they
all contain the exact same universal 'PROCPS_' prefix.
The origins for this are likely found in the desire to
avoid name clashes with other potential include files.
But with procps-ng newlib, we've probably gone way too
far. Did 'PROCPS_PIDS_TICS_SYSTEM' actually offer more
protection against clash than 'PIDS_TICS_SYSTEM' does?
I don't think so. Besides, no matter how big that name
becomes, one can never guarantee they'll never be some
clash. And, conversely, extremely short names will not
always create conflict. Of course, in either case when
some clash occurs, one can always #undef that problem.
Thus, this commit will eliminate that 'PROCPS_' prefix
making all of those enum identifiers a little shorter.
And, we'll still be well above some ridiculously short
(criminally short) names found in some common headers:
- - - - - - - - - - <term.h>
- 'tab', 'TTY', etc
- - - - - - - - - - - - - - - - <search.h>
- 'ENTER', ENTRY', 'FIND', etc
------------------------------------------------------
Finally, with this as a last of the wholesale changes,
we will have established the naming conventions below:
. only functions will begin with that 'procps_' prefix
. exposed structures begin with the module/header name
. item enumerators begin like structs, but capitalized
. other enumerators work exactly like item enumerators
. macros and constants begin just like the enumerators
------------------------------------------------------
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-21 10:30:00 +05:30
|
|
|
* my_stack = procps_diskstats_select(info, DISKSTATS_noop, num);
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
* ^~~~~~~~~~~~~~~~
|
|
|
|
*/
|
|
|
|
if (numitems < 1
|
library: removed all the 'PROCPS_' enumerator prefixes
Many of our item enumerator identifiers are very long,
especially in that <VMSTAT> module. Additionally, they
all contain the exact same universal 'PROCPS_' prefix.
The origins for this are likely found in the desire to
avoid name clashes with other potential include files.
But with procps-ng newlib, we've probably gone way too
far. Did 'PROCPS_PIDS_TICS_SYSTEM' actually offer more
protection against clash than 'PIDS_TICS_SYSTEM' does?
I don't think so. Besides, no matter how big that name
becomes, one can never guarantee they'll never be some
clash. And, conversely, extremely short names will not
always create conflict. Of course, in either case when
some clash occurs, one can always #undef that problem.
Thus, this commit will eliminate that 'PROCPS_' prefix
making all of those enum identifiers a little shorter.
And, we'll still be well above some ridiculously short
(criminally short) names found in some common headers:
- - - - - - - - - - <term.h>
- 'tab', 'TTY', etc
- - - - - - - - - - - - - - - - <search.h>
- 'ENTER', ENTRY', 'FIND', etc
------------------------------------------------------
Finally, with this as a last of the wholesale changes,
we will have established the naming conventions below:
. only functions will begin with that 'procps_' prefix
. exposed structures begin with the module/header name
. item enumerators begin like structs, but capitalized
. other enumerators work exactly like item enumerators
. macros and constants begin just like the enumerators
------------------------------------------------------
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-21 10:30:00 +05:30
|
|
|
|| (void *)items < (void *)(unsigned long)(2 * DISKSTATS_logical_end))
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
return -1;
|
|
|
|
|
|
|
|
for (i = 0; i < numitems; i++) {
|
|
|
|
// a diskstats_item is currently unsigned, but we'll protect our future
|
|
|
|
if (items[i] < 0)
|
|
|
|
return -1;
|
library: removed all the 'PROCPS_' enumerator prefixes
Many of our item enumerator identifiers are very long,
especially in that <VMSTAT> module. Additionally, they
all contain the exact same universal 'PROCPS_' prefix.
The origins for this are likely found in the desire to
avoid name clashes with other potential include files.
But with procps-ng newlib, we've probably gone way too
far. Did 'PROCPS_PIDS_TICS_SYSTEM' actually offer more
protection against clash than 'PIDS_TICS_SYSTEM' does?
I don't think so. Besides, no matter how big that name
becomes, one can never guarantee they'll never be some
clash. And, conversely, extremely short names will not
always create conflict. Of course, in either case when
some clash occurs, one can always #undef that problem.
Thus, this commit will eliminate that 'PROCPS_' prefix
making all of those enum identifiers a little shorter.
And, we'll still be well above some ridiculously short
(criminally short) names found in some common headers:
- - - - - - - - - - <term.h>
- 'tab', 'TTY', etc
- - - - - - - - - - - - - - - - <search.h>
- 'ENTER', ENTRY', 'FIND', etc
------------------------------------------------------
Finally, with this as a last of the wholesale changes,
we will have established the naming conventions below:
. only functions will begin with that 'procps_' prefix
. exposed structures begin with the module/header name
. item enumerators begin like structs, but capitalized
. other enumerators work exactly like item enumerators
. macros and constants begin just like the enumerators
------------------------------------------------------
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-21 10:30:00 +05:30
|
|
|
if (items[i] >= DISKSTATS_logical_end)
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
return -1;
|
|
|
|
}
|
2015-07-07 18:12:06 +05:30
|
|
|
|
|
|
|
return 0;
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
} // end: items_check_failed
|
|
|
|
|
2015-07-07 18:12:06 +05:30
|
|
|
|
|
|
|
/*
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
* read_diskstats_failed:
|
2015-07-07 18:12:06 +05:30
|
|
|
*
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
* @info: info structure created at procps_diskstats_new
|
2015-07-07 18:12:06 +05:30
|
|
|
*
|
|
|
|
* Read the data out of /proc/diskstats putting the information
|
|
|
|
* into the supplied info structure
|
|
|
|
*
|
|
|
|
* Returns: 0 on success, negative on error
|
|
|
|
*/
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
static int read_diskstats_failed (
|
2016-07-21 10:30:00 +05:30
|
|
|
struct diskstats_info *info)
|
2015-07-07 18:12:06 +05:30
|
|
|
{
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
static const char *fmtstr = "%d %d %" STRINGIFY(DISKSTATS_NAME_LEN) \
|
|
|
|
"s %lu %lu %lu %lu %lu %lu %lu %lu %lu %lu %lu";
|
|
|
|
char buf[DISKSTATS_LINE_LEN];
|
|
|
|
struct dev_node node;
|
|
|
|
int rc;
|
|
|
|
|
|
|
|
if (!info->diskstats_fp
|
|
|
|
&& (!(info->diskstats_fp = fopen(DISKSTATS_FILE, "r"))))
|
2015-07-07 18:12:06 +05:30
|
|
|
return -errno;
|
|
|
|
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
if (fseek(info->diskstats_fp, 0L, SEEK_SET) == -1)
|
2015-07-07 18:12:06 +05:30
|
|
|
return -errno;
|
|
|
|
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
info->old_stamp = info->new_stamp;
|
|
|
|
info->new_stamp = time(NULL);
|
|
|
|
|
|
|
|
while (fgets(buf, DISKSTATS_LINE_LEN, info->diskstats_fp)) {
|
|
|
|
// clear out the soon to be 'current'values
|
|
|
|
memset(&node, 0, sizeof(struct dev_node));
|
|
|
|
|
|
|
|
rc = sscanf(buf, fmtstr
|
|
|
|
, &node.major
|
|
|
|
, &node.minor
|
|
|
|
, &node.name[0]
|
|
|
|
, &node.new.reads
|
|
|
|
, &node.new.reads_merged
|
|
|
|
, &node.new.read_sectors
|
|
|
|
, &node.new.read_time
|
|
|
|
, &node.new.writes
|
|
|
|
, &node.new.writes_merged
|
|
|
|
, &node.new.write_sectors
|
|
|
|
, &node.new.write_time
|
|
|
|
, &node.new.io_inprogress
|
|
|
|
, &node.new.io_time
|
|
|
|
, &node.new.io_wtime);
|
|
|
|
|
|
|
|
if (rc != 14) {
|
2015-07-07 18:12:06 +05:30
|
|
|
if (errno != 0)
|
|
|
|
return -errno;
|
2016-07-24 10:30:00 +05:30
|
|
|
return -EIO;
|
2015-07-07 18:12:06 +05:30
|
|
|
}
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
node.stamped = info->new_stamp;
|
|
|
|
if ((rc = node_update(info, &node)))
|
|
|
|
return rc;
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
} //end: read_diskstats_failed
|
|
|
|
|
|
|
|
|
|
|
|
static struct stacks_extent *stacks_alloc (
|
|
|
|
struct ext_support *this,
|
|
|
|
int maxstacks)
|
|
|
|
{
|
|
|
|
struct stacks_extent *p_blob;
|
|
|
|
struct diskstats_stack **p_vect;
|
|
|
|
struct diskstats_stack *p_head;
|
|
|
|
size_t vect_size, head_size, list_size, blob_size;
|
|
|
|
void *v_head, *v_list;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
if (this == NULL || this->items == NULL)
|
|
|
|
return NULL;
|
|
|
|
if (maxstacks < 1)
|
|
|
|
return NULL;
|
|
|
|
|
2016-07-24 10:30:00 +05:30
|
|
|
vect_size = sizeof(void *) * maxstacks; // size of the addr vectors |
|
|
|
|
vect_size += sizeof(void *); // plus NULL addr delimiter |
|
|
|
|
head_size = sizeof(struct diskstats_stack); // size of that head struct |
|
|
|
|
list_size = sizeof(struct diskstats_result) * this->numitems; // any single results stack |
|
|
|
|
blob_size = sizeof(struct stacks_extent); // the extent anchor itself |
|
|
|
|
blob_size += vect_size; // plus room for addr vects |
|
|
|
|
blob_size += head_size * maxstacks; // plus room for head thing |
|
|
|
|
blob_size += list_size * maxstacks; // plus room for our stacks |
|
|
|
|
|
|
|
|
/* note: all of our memory is allocated in one single blob, facilitating some later free(). |
|
|
|
|
as a minimum, it's important that all of those result structs themselves always be |
|
|
|
|
contiguous within every stack since they will be accessed via a relative position. | */
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
if (NULL == (p_blob = calloc(1, blob_size)))
|
|
|
|
return NULL;
|
|
|
|
|
2016-07-24 10:30:00 +05:30
|
|
|
p_blob->next = this->extents; // push this extent onto... |
|
|
|
|
this->extents = p_blob; // ...some existing extents |
|
|
|
|
p_vect = (void *)p_blob + sizeof(struct stacks_extent); // prime our vector pointer |
|
|
|
|
p_blob->stacks = p_vect; // set actual vectors start |
|
|
|
|
v_head = (void *)p_vect + vect_size; // prime head pointer start |
|
|
|
|
v_list = v_head + (head_size * maxstacks); // prime our stacks pointer |
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
|
|
|
|
for (i = 0; i < maxstacks; i++) {
|
|
|
|
p_head = (struct diskstats_stack *)v_head;
|
|
|
|
p_head->head = itemize_stack((struct diskstats_result *)v_list, this->numitems, this->items);
|
|
|
|
p_blob->stacks[i] = p_head;
|
|
|
|
v_list += list_size;
|
|
|
|
v_head += head_size;
|
|
|
|
}
|
|
|
|
p_blob->ext_numstacks = maxstacks;
|
|
|
|
return p_blob;
|
|
|
|
} // end: stacks_alloc
|
|
|
|
|
|
|
|
|
|
|
|
static int stacks_fetch (
|
2016-07-21 10:30:00 +05:30
|
|
|
struct diskstats_info *info)
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
{
|
|
|
|
#define n_alloc info->fetch.n_alloc
|
|
|
|
#define n_inuse info->fetch.n_inuse
|
|
|
|
#define n_saved info->fetch.n_alloc_save
|
|
|
|
struct stacks_extent *ext;
|
|
|
|
struct dev_node *node;
|
|
|
|
|
|
|
|
// initialize stuff -----------------------------------
|
|
|
|
if (!info->fetch.anchor) {
|
|
|
|
if (!(info->fetch.anchor = calloc(sizeof(void *), STACKS_INCR)))
|
|
|
|
return -ENOMEM;
|
|
|
|
n_alloc = STACKS_INCR;
|
|
|
|
}
|
|
|
|
if (!info->fetch_ext.extents) {
|
|
|
|
if (!(ext = stacks_alloc(&info->fetch_ext, n_alloc)))
|
|
|
|
return -ENOMEM;
|
|
|
|
memset(info->fetch.anchor, 0, sizeof(void *) * n_alloc);
|
|
|
|
memcpy(info->fetch.anchor, ext->stacks, sizeof(void *) * n_alloc);
|
|
|
|
itemize_stacks_all(&info->fetch_ext);
|
|
|
|
}
|
|
|
|
cleanup_stacks_all(&info->fetch_ext);
|
|
|
|
|
|
|
|
// iterate stuff --------------------------------------
|
|
|
|
n_inuse = 0;
|
|
|
|
node = info->nodes;
|
|
|
|
while (node) {
|
|
|
|
if (!(n_inuse < n_alloc)) {
|
|
|
|
n_alloc += STACKS_INCR;
|
|
|
|
if ((!(info->fetch.anchor = realloc(info->fetch.anchor, sizeof(void *) * n_alloc)))
|
|
|
|
|| (!(ext = stacks_alloc(&info->fetch_ext, STACKS_INCR))))
|
|
|
|
return -1;
|
|
|
|
memcpy(info->fetch.anchor + n_inuse, ext->stacks, sizeof(void *) * STACKS_INCR);
|
2015-07-07 18:12:06 +05:30
|
|
|
}
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
assign_results(info->fetch.anchor[n_inuse], node);
|
|
|
|
++n_inuse;
|
|
|
|
node = node->next;
|
|
|
|
}
|
|
|
|
|
|
|
|
// finalize stuff -------------------------------------
|
|
|
|
/* note: we go to this trouble of maintaining a duplicate of the consolidated |
|
|
|
|
extent stacks addresses represented as our 'anchor' since these ptrs |
|
|
|
|
are exposed to a user (um, not that we don't trust 'em or anything). |
|
|
|
|
plus, we can NULL delimit these ptrs which we couldn't do otherwise. | */
|
|
|
|
if (n_saved < n_inuse + 1) {
|
|
|
|
n_saved = n_inuse + 1;
|
|
|
|
if (!(info->fetch.results.stacks = realloc(info->fetch.results.stacks, sizeof(void *) * n_saved)))
|
|
|
|
return -ENOMEM;
|
|
|
|
}
|
|
|
|
memcpy(info->fetch.results.stacks, info->fetch.anchor, sizeof(void *) * n_inuse);
|
|
|
|
info->fetch.results.stacks[n_inuse] = NULL;
|
|
|
|
info->fetch.results.total = n_inuse;
|
|
|
|
|
|
|
|
return n_inuse;
|
|
|
|
#undef n_alloc
|
|
|
|
#undef n_inuse
|
|
|
|
#undef n_saved
|
|
|
|
} // end: stacks_fetch
|
|
|
|
|
|
|
|
|
|
|
|
static int stacks_reconfig_maybe (
|
|
|
|
struct ext_support *this,
|
|
|
|
enum diskstats_item *items,
|
|
|
|
int numitems)
|
|
|
|
{
|
|
|
|
if (items_check_failed(this, items, numitems))
|
|
|
|
return -EINVAL;
|
|
|
|
/* is this the first time or have things changed since we were last called?
|
|
|
|
if so, gotta' redo all of our stacks stuff ... */
|
|
|
|
if (this->numitems != numitems + 1
|
|
|
|
|| memcmp(this->items, items, sizeof(enum diskstats_item) * numitems)) {
|
library: removed all the 'PROCPS_' enumerator prefixes
Many of our item enumerator identifiers are very long,
especially in that <VMSTAT> module. Additionally, they
all contain the exact same universal 'PROCPS_' prefix.
The origins for this are likely found in the desire to
avoid name clashes with other potential include files.
But with procps-ng newlib, we've probably gone way too
far. Did 'PROCPS_PIDS_TICS_SYSTEM' actually offer more
protection against clash than 'PIDS_TICS_SYSTEM' does?
I don't think so. Besides, no matter how big that name
becomes, one can never guarantee they'll never be some
clash. And, conversely, extremely short names will not
always create conflict. Of course, in either case when
some clash occurs, one can always #undef that problem.
Thus, this commit will eliminate that 'PROCPS_' prefix
making all of those enum identifiers a little shorter.
And, we'll still be well above some ridiculously short
(criminally short) names found in some common headers:
- - - - - - - - - - <term.h>
- 'tab', 'TTY', etc
- - - - - - - - - - - - - - - - <search.h>
- 'ENTER', ENTRY', 'FIND', etc
------------------------------------------------------
Finally, with this as a last of the wholesale changes,
we will have established the naming conventions below:
. only functions will begin with that 'procps_' prefix
. exposed structures begin with the module/header name
. item enumerators begin like structs, but capitalized
. other enumerators work exactly like item enumerators
. macros and constants begin just like the enumerators
------------------------------------------------------
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-21 10:30:00 +05:30
|
|
|
// allow for our DISKSTATS_logical_end
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
if (!(this->items = realloc(this->items, sizeof(enum diskstats_item) * (numitems + 1))))
|
|
|
|
return -ENOMEM;
|
|
|
|
memcpy(this->items, items, sizeof(enum diskstats_item) * numitems);
|
library: removed all the 'PROCPS_' enumerator prefixes
Many of our item enumerator identifiers are very long,
especially in that <VMSTAT> module. Additionally, they
all contain the exact same universal 'PROCPS_' prefix.
The origins for this are likely found in the desire to
avoid name clashes with other potential include files.
But with procps-ng newlib, we've probably gone way too
far. Did 'PROCPS_PIDS_TICS_SYSTEM' actually offer more
protection against clash than 'PIDS_TICS_SYSTEM' does?
I don't think so. Besides, no matter how big that name
becomes, one can never guarantee they'll never be some
clash. And, conversely, extremely short names will not
always create conflict. Of course, in either case when
some clash occurs, one can always #undef that problem.
Thus, this commit will eliminate that 'PROCPS_' prefix
making all of those enum identifiers a little shorter.
And, we'll still be well above some ridiculously short
(criminally short) names found in some common headers:
- - - - - - - - - - <term.h>
- 'tab', 'TTY', etc
- - - - - - - - - - - - - - - - <search.h>
- 'ENTER', ENTRY', 'FIND', etc
------------------------------------------------------
Finally, with this as a last of the wholesale changes,
we will have established the naming conventions below:
. only functions will begin with that 'procps_' prefix
. exposed structures begin with the module/header name
. item enumerators begin like structs, but capitalized
. other enumerators work exactly like item enumerators
. macros and constants begin just like the enumerators
------------------------------------------------------
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-21 10:30:00 +05:30
|
|
|
this->items[numitems] = DISKSTATS_logical_end;
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
this->numitems = numitems + 1;
|
|
|
|
if (this->extents)
|
|
|
|
extents_free_all(this);
|
|
|
|
return 1;
|
2015-07-07 18:12:06 +05:30
|
|
|
}
|
|
|
|
return 0;
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
} // end: stacks_reconfig_maybe
|
|
|
|
|
|
|
|
|
|
|
|
// ___ Public Functions |||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|
|
|
|
|
|
// --- standard required functions --------------------------------------------
|
2015-07-07 18:12:06 +05:30
|
|
|
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
/*
|
|
|
|
* procps_diskstats_new():
|
|
|
|
*
|
|
|
|
* @info: location of returned new structure
|
|
|
|
*
|
|
|
|
* Returns: < 0 on failure, 0 on success along with
|
|
|
|
* a pointer to a new context struct
|
|
|
|
*/
|
|
|
|
PROCPS_EXPORT int procps_diskstats_new (
|
2016-07-21 10:30:00 +05:30
|
|
|
struct diskstats_info **info)
|
2016-05-01 11:45:49 +05:30
|
|
|
{
|
2016-07-21 10:30:00 +05:30
|
|
|
struct diskstats_info *p;
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
int rc;
|
|
|
|
|
2016-05-01 11:45:49 +05:30
|
|
|
if (info == NULL)
|
|
|
|
return -EINVAL;
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
|
2016-07-21 10:30:00 +05:30
|
|
|
if (!(p = calloc(1, sizeof(struct diskstats_info))))
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
p->refcount = 1;
|
|
|
|
|
|
|
|
/* do a priming read here for the following potential benefits: |
|
|
|
|
1) ensure there will be no problems with subsequent access |
|
|
|
|
2) make delta results potentially useful, even if 1st time | */
|
|
|
|
if ((rc = read_diskstats_failed(p))) {
|
|
|
|
procps_diskstats_unref(&p);
|
|
|
|
return rc;
|
|
|
|
}
|
|
|
|
|
|
|
|
*info = p;
|
|
|
|
return 0;
|
|
|
|
} // end: procps_diskstats_new
|
|
|
|
|
|
|
|
|
|
|
|
PROCPS_EXPORT int procps_diskstats_ref (
|
2016-07-21 10:30:00 +05:30
|
|
|
struct diskstats_info *info)
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
{
|
|
|
|
if (info == NULL)
|
|
|
|
return -EINVAL;
|
|
|
|
|
2016-05-01 11:45:49 +05:30
|
|
|
info->refcount++;
|
|
|
|
return info->refcount;
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
} // end: procps_diskstats_ref
|
|
|
|
|
2016-05-01 11:45:49 +05:30
|
|
|
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
PROCPS_EXPORT int procps_diskstats_unref (
|
2016-07-21 10:30:00 +05:30
|
|
|
struct diskstats_info **info)
|
2016-05-01 11:45:49 +05:30
|
|
|
{
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
struct dev_node *node;
|
|
|
|
|
2016-05-01 11:45:49 +05:30
|
|
|
if (info == NULL || *info == NULL)
|
|
|
|
return -EINVAL;
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
|
2016-05-01 11:45:49 +05:30
|
|
|
(*info)->refcount--;
|
|
|
|
if ((*info)->refcount == 0) {
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
if ((*info)->diskstats_fp) {
|
|
|
|
fclose((*info)->diskstats_fp);
|
|
|
|
(*info)->diskstats_fp = NULL;
|
|
|
|
}
|
|
|
|
node = (*info)->nodes;
|
|
|
|
while (node) {
|
|
|
|
struct dev_node *p = node;
|
|
|
|
node = p->next;
|
|
|
|
free(p);
|
|
|
|
}
|
|
|
|
if ((*info)->select_ext.extents)
|
|
|
|
extents_free_all((&(*info)->select_ext));
|
|
|
|
if ((*info)->select_ext.items)
|
|
|
|
free((*info)->select_ext.items);
|
|
|
|
|
|
|
|
if ((*info)->fetch.anchor)
|
|
|
|
free((*info)->fetch.anchor);
|
|
|
|
if ((*info)->fetch.results.stacks)
|
|
|
|
free((*info)->fetch.results.stacks);
|
|
|
|
|
|
|
|
if ((*info)->fetch_ext.extents)
|
|
|
|
extents_free_all(&(*info)->fetch_ext);
|
|
|
|
if ((*info)->fetch_ext.items)
|
|
|
|
free((*info)->fetch_ext.items);
|
|
|
|
|
2016-05-01 11:45:49 +05:30
|
|
|
free(*info);
|
|
|
|
*info = NULL;
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
|
2016-05-01 11:45:49 +05:30
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
return (*info)->refcount;
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
} // end: procps_diskstats_unref
|
2016-05-01 11:45:49 +05:30
|
|
|
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
|
|
|
|
// --- variable interface functions -------------------------------------------
|
|
|
|
|
|
|
|
PROCPS_EXPORT struct diskstats_result *procps_diskstats_get (
|
2016-07-21 10:30:00 +05:30
|
|
|
struct diskstats_info *info,
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
const char *name,
|
|
|
|
enum diskstats_item item)
|
|
|
|
{
|
|
|
|
struct dev_node *node;
|
|
|
|
time_t cur_secs;
|
|
|
|
|
|
|
|
if (info == NULL)
|
|
|
|
return NULL;
|
library: removed all the 'PROCPS_' enumerator prefixes
Many of our item enumerator identifiers are very long,
especially in that <VMSTAT> module. Additionally, they
all contain the exact same universal 'PROCPS_' prefix.
The origins for this are likely found in the desire to
avoid name clashes with other potential include files.
But with procps-ng newlib, we've probably gone way too
far. Did 'PROCPS_PIDS_TICS_SYSTEM' actually offer more
protection against clash than 'PIDS_TICS_SYSTEM' does?
I don't think so. Besides, no matter how big that name
becomes, one can never guarantee they'll never be some
clash. And, conversely, extremely short names will not
always create conflict. Of course, in either case when
some clash occurs, one can always #undef that problem.
Thus, this commit will eliminate that 'PROCPS_' prefix
making all of those enum identifiers a little shorter.
And, we'll still be well above some ridiculously short
(criminally short) names found in some common headers:
- - - - - - - - - - <term.h>
- 'tab', 'TTY', etc
- - - - - - - - - - - - - - - - <search.h>
- 'ENTER', ENTRY', 'FIND', etc
------------------------------------------------------
Finally, with this as a last of the wholesale changes,
we will have established the naming conventions below:
. only functions will begin with that 'procps_' prefix
. exposed structures begin with the module/header name
. item enumerators begin like structs, but capitalized
. other enumerators work exactly like item enumerators
. macros and constants begin just like the enumerators
------------------------------------------------------
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-21 10:30:00 +05:30
|
|
|
if (item < 0 || item >= DISKSTATS_logical_end)
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
return NULL;
|
|
|
|
|
|
|
|
/* we will NOT read the diskstat file with every call - rather, we'll offer
|
|
|
|
a granularity of 1 second between reads ... */
|
|
|
|
cur_secs = time(NULL);
|
|
|
|
if (1 <= cur_secs - info->new_stamp) {
|
|
|
|
if (read_diskstats_failed(info))
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
info->get_this.item = item;
|
|
|
|
// with 'get', we must NOT honor the usual 'noop' guarantee
|
library: removed all the 'PROCPS_' enumerator prefixes
Many of our item enumerator identifiers are very long,
especially in that <VMSTAT> module. Additionally, they
all contain the exact same universal 'PROCPS_' prefix.
The origins for this are likely found in the desire to
avoid name clashes with other potential include files.
But with procps-ng newlib, we've probably gone way too
far. Did 'PROCPS_PIDS_TICS_SYSTEM' actually offer more
protection against clash than 'PIDS_TICS_SYSTEM' does?
I don't think so. Besides, no matter how big that name
becomes, one can never guarantee they'll never be some
clash. And, conversely, extremely short names will not
always create conflict. Of course, in either case when
some clash occurs, one can always #undef that problem.
Thus, this commit will eliminate that 'PROCPS_' prefix
making all of those enum identifiers a little shorter.
And, we'll still be well above some ridiculously short
(criminally short) names found in some common headers:
- - - - - - - - - - <term.h>
- 'tab', 'TTY', etc
- - - - - - - - - - - - - - - - <search.h>
- 'ENTER', ENTRY', 'FIND', etc
------------------------------------------------------
Finally, with this as a last of the wholesale changes,
we will have established the naming conventions below:
. only functions will begin with that 'procps_' prefix
. exposed structures begin with the module/header name
. item enumerators begin like structs, but capitalized
. other enumerators work exactly like item enumerators
. macros and constants begin just like the enumerators
------------------------------------------------------
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-21 10:30:00 +05:30
|
|
|
// if (item > DISKSTATS_noop)
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
info->get_this.result.ul_int = 0;
|
|
|
|
|
|
|
|
if (!(node = node_get(info, name)))
|
|
|
|
return NULL;
|
|
|
|
Item_table[item].setsfunc(&info->get_this, node);
|
|
|
|
|
|
|
|
return &info->get_this;
|
|
|
|
} // end: procps_diskstats_get
|
|
|
|
|
|
|
|
|
|
|
|
/* procps_diskstats_reap():
|
2015-07-07 18:12:06 +05:30
|
|
|
*
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
* Harvest all the requested disks information providing
|
|
|
|
* the result stacks along with the total number of harvested.
|
2015-07-07 18:12:06 +05:30
|
|
|
*
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
* Returns: pointer to a diskstats_reap struct on success, NULL on error.
|
2015-07-07 18:12:06 +05:30
|
|
|
*/
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
PROCPS_EXPORT struct diskstats_reap *procps_diskstats_reap (
|
2016-07-21 10:30:00 +05:30
|
|
|
struct diskstats_info *info,
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
enum diskstats_item *items,
|
|
|
|
int numitems)
|
2015-07-07 18:12:06 +05:30
|
|
|
{
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
if (info == NULL || items == NULL)
|
|
|
|
return NULL;
|
2015-07-07 18:12:06 +05:30
|
|
|
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
if (0 > stacks_reconfig_maybe(&info->fetch_ext, items, numitems))
|
|
|
|
return NULL;
|
2015-07-07 18:12:06 +05:30
|
|
|
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
if (info->fetch_ext.dirty_stacks)
|
|
|
|
cleanup_stacks_all(&info->fetch_ext);
|
|
|
|
|
|
|
|
if (read_diskstats_failed(info))
|
|
|
|
return NULL;
|
|
|
|
stacks_fetch(info);
|
|
|
|
info->fetch_ext.dirty_stacks = 1;
|
|
|
|
|
|
|
|
return &info->fetch.results;
|
|
|
|
} // end: procps_diskstats_reap
|
|
|
|
|
|
|
|
|
|
|
|
/* procps_diskstats_select():
|
2015-07-07 18:12:06 +05:30
|
|
|
*
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
* Obtain all the requested disk/partition information then return
|
|
|
|
* it in a single library provided results stack.
|
2015-07-07 18:12:06 +05:30
|
|
|
*
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
* Returns: pointer to a diskstats_stack struct on success, NULL on error.
|
2015-07-07 18:12:06 +05:30
|
|
|
*/
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
PROCPS_EXPORT struct diskstats_stack *procps_diskstats_select (
|
2016-07-21 10:30:00 +05:30
|
|
|
struct diskstats_info *info,
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
const char *name,
|
|
|
|
enum diskstats_item *items,
|
|
|
|
int numitems)
|
2015-07-07 18:12:06 +05:30
|
|
|
{
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
struct dev_node *node;
|
2015-07-07 18:12:06 +05:30
|
|
|
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
if (info == NULL || items == NULL)
|
|
|
|
return NULL;
|
2015-07-07 18:12:06 +05:30
|
|
|
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
if (0 > stacks_reconfig_maybe(&info->select_ext, items, numitems))
|
|
|
|
return NULL;
|
2015-07-07 18:12:06 +05:30
|
|
|
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
if (!info->select_ext.extents
|
|
|
|
&& !(stacks_alloc(&info->select_ext, 1)))
|
|
|
|
return NULL;
|
2015-07-07 18:12:06 +05:30
|
|
|
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
if (info->select_ext.dirty_stacks)
|
|
|
|
cleanup_stacks_all(&info->select_ext);
|
|
|
|
|
|
|
|
if (read_diskstats_failed(info))
|
|
|
|
return NULL;
|
|
|
|
if (!(node = node_get(info, name)))
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
assign_results(info->select_ext.extents->stacks[0], node);
|
|
|
|
info->select_ext.dirty_stacks = 1;
|
|
|
|
|
|
|
|
return info->select_ext.extents->stacks[0];
|
|
|
|
} // end: procps_diskstats_select
|
2015-07-07 18:12:06 +05:30
|
|
|
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
|
|
|
|
/*
|
|
|
|
* procps_diskstats_sort():
|
|
|
|
*
|
|
|
|
* Sort stacks anchored in the passed stack pointers array
|
|
|
|
* based on the designated sort enumerator and specified order.
|
|
|
|
*
|
|
|
|
* Returns those same addresses sorted.
|
|
|
|
*
|
|
|
|
* Note: all of the stacks must be homogeneous (of equal length and content).
|
|
|
|
*/
|
|
|
|
PROCPS_EXPORT struct diskstats_stack **procps_diskstats_sort (
|
2016-07-21 10:30:00 +05:30
|
|
|
struct diskstats_info *info,
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
struct diskstats_stack *stacks[],
|
|
|
|
int numstacked,
|
|
|
|
enum diskstats_item sortitem,
|
|
|
|
enum diskstats_sort_order order)
|
2015-07-07 18:12:06 +05:30
|
|
|
{
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
struct diskstats_result *p;
|
|
|
|
struct sort_parms parms;
|
|
|
|
int offset;
|
|
|
|
|
|
|
|
if (info == NULL || stacks == NULL)
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
// a diskstats_item is currently unsigned, but we'll protect our future
|
library: removed all the 'PROCPS_' enumerator prefixes
Many of our item enumerator identifiers are very long,
especially in that <VMSTAT> module. Additionally, they
all contain the exact same universal 'PROCPS_' prefix.
The origins for this are likely found in the desire to
avoid name clashes with other potential include files.
But with procps-ng newlib, we've probably gone way too
far. Did 'PROCPS_PIDS_TICS_SYSTEM' actually offer more
protection against clash than 'PIDS_TICS_SYSTEM' does?
I don't think so. Besides, no matter how big that name
becomes, one can never guarantee they'll never be some
clash. And, conversely, extremely short names will not
always create conflict. Of course, in either case when
some clash occurs, one can always #undef that problem.
Thus, this commit will eliminate that 'PROCPS_' prefix
making all of those enum identifiers a little shorter.
And, we'll still be well above some ridiculously short
(criminally short) names found in some common headers:
- - - - - - - - - - <term.h>
- 'tab', 'TTY', etc
- - - - - - - - - - - - - - - - <search.h>
- 'ENTER', ENTRY', 'FIND', etc
------------------------------------------------------
Finally, with this as a last of the wholesale changes,
we will have established the naming conventions below:
. only functions will begin with that 'procps_' prefix
. exposed structures begin with the module/header name
. item enumerators begin like structs, but capitalized
. other enumerators work exactly like item enumerators
. macros and constants begin just like the enumerators
------------------------------------------------------
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-21 10:30:00 +05:30
|
|
|
if (sortitem < 0 || sortitem >= DISKSTATS_logical_end)
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
return NULL;
|
library: removed all the 'PROCPS_' enumerator prefixes
Many of our item enumerator identifiers are very long,
especially in that <VMSTAT> module. Additionally, they
all contain the exact same universal 'PROCPS_' prefix.
The origins for this are likely found in the desire to
avoid name clashes with other potential include files.
But with procps-ng newlib, we've probably gone way too
far. Did 'PROCPS_PIDS_TICS_SYSTEM' actually offer more
protection against clash than 'PIDS_TICS_SYSTEM' does?
I don't think so. Besides, no matter how big that name
becomes, one can never guarantee they'll never be some
clash. And, conversely, extremely short names will not
always create conflict. Of course, in either case when
some clash occurs, one can always #undef that problem.
Thus, this commit will eliminate that 'PROCPS_' prefix
making all of those enum identifiers a little shorter.
And, we'll still be well above some ridiculously short
(criminally short) names found in some common headers:
- - - - - - - - - - <term.h>
- 'tab', 'TTY', etc
- - - - - - - - - - - - - - - - <search.h>
- 'ENTER', ENTRY', 'FIND', etc
------------------------------------------------------
Finally, with this as a last of the wholesale changes,
we will have established the naming conventions below:
. only functions will begin with that 'procps_' prefix
. exposed structures begin with the module/header name
. item enumerators begin like structs, but capitalized
. other enumerators work exactly like item enumerators
. macros and constants begin just like the enumerators
------------------------------------------------------
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-21 10:30:00 +05:30
|
|
|
if (order != DISKSTATS_SORT_ASCEND && order != DISKSTATS_SORT_DESCEND)
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
return NULL;
|
|
|
|
if (numstacked < 2)
|
|
|
|
return stacks;
|
|
|
|
|
|
|
|
offset = 0;
|
|
|
|
p = stacks[0]->head;
|
|
|
|
for (;;) {
|
|
|
|
if (p->item == sortitem)
|
|
|
|
break;
|
|
|
|
++offset;
|
2016-07-24 10:30:00 +05:30
|
|
|
if (p->item >= DISKSTATS_logical_end)
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
return NULL;
|
|
|
|
++p;
|
|
|
|
}
|
|
|
|
parms.offset = offset;
|
|
|
|
parms.order = order;
|
2015-07-07 18:12:06 +05:30
|
|
|
|
library: normalize/standardize an i/f, <DISKSTATS> api
This patch will bring this interface up to our 3rd gen
standards. The following summarizes the major changes.
* New delta provisions have been added to most fields.
There are, of course, some fields for which a delta is
inappropriate. They include the identifying items such
as name, type, major and minor. Plus the io_inprogress
field which already acts, in effect, as a delta value.
* To provide delta support, dev_node historical values
have become persistent. By the same token, the library
must provide for future removal of disks/partitions. A
timestamp is used to detect 'stale' data which will be
deleted so as not to satisfy some get, select or reap.
* Such persistent support is provided by a linked list
which, by default, grows from the bottom down so as to
maintain compatibility with the /proc/diskstats order.
Initially, I was tempted to use the GNU tsearch (tree)
provisions until I discovered the overhead of building
that tree plus costs of a subsequent 'twalk'. Besides,
walking such a tree means retrieval order would differ
from an order required/expected by the vmstat program.
* The '/sys/block' directory is no longer scanned with
every refresh cycle. Rather, it's only accessed when a
node is first encountered. Then, that node's 'type' is
persistent for its lifetime like several other fields.
* A sort provision was included, at virtually no cost,
even though such a provision was not currently needed.
Signed-off-by: Jim Warner <james.warner@comcast.net>
2016-07-20 10:30:00 +05:30
|
|
|
qsort_r(stacks, numstacked, sizeof(void *), (QSR_t)Item_table[p->item].sortfunc, &parms);
|
|
|
|
return stacks;
|
|
|
|
} // end: procps_diskstats_sort
|