procps/doc/procps_pids.3

.\" (C) Copyright 2020-2022 Jim Warner <warnerjc@comcast.net>
.\"
.\" %%%LICENSE_START(LGPL_2.1+)
.\" This manual 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 manual 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
.\" %%%LICENSE_END
.\"
.TH PROCPS_PIDS 3 "January 2022" "libproc-2"
.\" Please adjust this date whenever revising the manpage.
.\"
.nh
.SH NAME
procps_pids \- API to access process information in the /proc filesystem

.SH SYNOPSIS
.nf
#include <procps/pids.h>

.BI "int procps_pids_new   (struct pids_info **" info ", enum pids_item *" items ", int " numitems );
.BI "int procps_pids_ref   (struct pids_info  *" info );
.BI "int procps_pids_unref (struct pids_info **" info );


.BI "struct pids_stack *procps_pids_get ("
.BI "    struct pids_info *" info ,
.BI "    enum pids_fetch_type " which );

.BI "struct pids_fetch *procps_pids_reap ("
.BI "    struct pids_info *" info ,
.BI "    enum pids_fetch_type " which );

.BI "struct pids_fetch *procps_pids_select ("
.BI "    struct pids_info *" info ,
.BI "    unsigned *" these ,
.BI "    int " numthese ,
.BI "    enum pids_select_type " which );

.BI "struct pids_stack **procps_pids_sort ("
.BI "    struct pids_info *" info ,
.BI "    struct pids_stack *" stacks[] ,
.BI "    int " numstacked ,
.BI "    enum pids_item " sortitem ,
.BI "    enum pids_sort_order " order );


.BI "int procps_pids_reset ("
.BI "    struct pids_info *" info ,
.BI "    enum pids_item *" newitems ,
.BI "    int " newnumitems );

.BI "struct pids_stack *fatal_proc_unmounted ("
.BI "    struct pids_info *" info ,
.BI "    int " return_self );

.fi

Link with \fI\-lproc-2\fP.

.SH DESCRIPTION
.SS Overview
Central to this interface is a simple `result'
structure reflecting an `item' plus its value (in a union
with standard C language types as members).
All `result' structures are automatically allocated and
provided by the library.

By specifying an array of `items', these structures can be
organized as a `stack', potentially yielding many results
with a single function call.
Thus, a `stack' can be viewed as a variable length record
whose content and order is determined solely by the user.

As part of this interface there are two unique enumerators.
The `noop' and `extra' items exist to hold user values.
They are never set by the library, but the `extra'
result will be zeroed with each library interaction.

The pids.h file will be an essential document during
user program development.
There you will find available items, their return type
(the `result' struct member name) and the source for such values.
Additional enumerators and structures are also documented there.

.SS Usage
The following would be a typical sequence of calls to
this interface.

.nf
.RB "1. " fatal_proc_unmounted()
.RB "2. " procps_pids_new()
.RB "3. " procps_pids_get() ", " procps_pids_reap() " or " procps_pids_select()
.RB "4. " procps_pids_unref()
.fi

The \fBget\fR function is an iterator for successive PIDs/TIDs,
returning those `items' previously identified via \fBnew\fR
or \fBreset\fR.

Two functions support unpredictable variable outcomes.
The \fBreap\fR function gathers data for all processes while
the \fBselect\fR function deals with specific PIDs or UIDs.
Both can return multiple `stacks' each containing multiple `result'
structures.
Optionally, a user may choose to \fBsort\fR such results

To exploit any `stack', and access individual `result' structures,
a \fIrelative_enum\fR is required as shown in the \fBVAL\fR macro
defined in the header file.
Such values could be hard coded as: 0 through numitems-1.
However, this need is typically satisfied by creating your own
enumerators corresponding to the order of the `items' array.

.SS Caveats
The <pids> API differs from others in that those items
of interest must be provided at \fBnew\fR or \fBreset\fR time,
the latter being unique to this API.
If either the \fIitems\fR or \fInumitems\fR parameter is zero at
\fBnew\fR time, then \fBreset\fR becomes mandatory before
issuing any other call.

For the \fBnew\fR and \fBunref\fR functions, the address of an \fIinfo\fR
struct pointer must be supplied.
With \fBnew\fR it must have been initialized to NULL.
With \fBunref\fR it will be reset to NULL if the reference count reaches zero.

The \fBget\fR and \fBreap\fR functions use the \fIwhich\fR parameter
to specify whether just tasks or both tasks and threads are to be fetched.

The \fBselect\fR function requires an array of PIDs or UIDs as
\fIthese\fR along with \fInumthese\fR to identify which processes
are to be fetched.
This function then operates as a subset of \fBreap\fR.

When using the \fBsort\fR function, the parameters \fIstacks\fR and
\fInumstacked\fR would normally be those returned in the `pids_fetch'
structure.

Lastly, a \fBfatal_proc_unmounted\fR function may be called before
any other function to ensure that the /proc/ directory is mounted.
As such, the \fIinfo\fR parameter would be NULL and the
\fIreturn_self\fR parameter zero.
If, however, some items are desired for the issuing program (a
\fIreturn_self\fR other than zero) then the \fBnew\fR call must precede
it to identify the \fIitems\fR and obtain the required \fIinfo\fR pointer.

.SH RETURN VALUE
.SS Functions Returning an `int'
An error will be indicated by a negative number that
is always the inverse of some well known errno.h value.

Success is indicated by a zero return value.
However, the \fBref\fR and \fBunref\fR functions return
the current \fIinfo\fR structure reference count.

.SS Functions Returning an `address'
An error will be indicated by a NULL return pointer
with the reason found in the formal errno value.

Success is indicated by a pointer to the named structure.
However, if one survives the \fBfatal_proc_unmounted\fR call,
NULL is always returned when \fIreturn_self\fR is zero.

.SH DEBUGGING
To aid in program development, there are two procps-ng provisions
that can be exploited.

The first is a supplied file named `libproc.supp' which may be
useful when developing a \fImulti-threaded\fR application.
When used with the valgrind `--suppressions=' option, warnings
associated with the procps library itself are avoided.

Such warnings arise because the library handles heap based
allocations in a thread-safe manner.
A \fIsingle-threaded\fR application will not receive those warnings.

The second provision can help ensure `result' member references
agree with library expectations.
It assumes that a supplied macro in the header file is
used to access the `result' value.

This feature can be activated through either of the following
methods and any discrepancies will be written to \fBstderr\fR.

.IP 1) 3
Add CFLAGS='-DXTRA_PROCPS_DEBUG' to any other ./configure
options your project may employ.

.IP 2) 3
Add #include <procps/xtra-procps-debug.h> to any program
\fIafter\fR the #include <procps/pids.h>.

.PP
This verification feature incurs substantial overhead.
Therefore, it is important that it \fInot\fR be activated
for a production/release build.

.SH SEE ALSO
.BR procps (3)
.BR procps_misc (3)
.BR proc (5).