oxmc/libgtop
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

New directory. Updated documentation.

Browse Source 
1998-08-12  Martin Baulig  <martin@home-of-linux.org>

	* features: New directory.
	* gnome-hackers.sgml: Updated documentation.
This commit is contained in:
Martin Baulig
1998-08-12 16:25:48 +00:00
committed by Martin Baulig
parent 2356852a8c
commit b05cfacb3d
7 changed files with 364 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

153
doc/features/proclist.sgml Normal file
View File

@@ -0,0 +1,153 @@
<para>
This functions takes the following additional parameters:
<variablelist>
<varlistentry>
<term><parameter>which</parameter></term>
<listitem>
<para>
You can use the following constants (defined in
<filename>glibtop/proclist.h</filename>) to tell
<function>glibtop_get_proclist ()</function> which processes
to fetch:
<variablelist>
<varlistentry>
<term><parameter>GLIBTOP_KERN_PROC_ALL</parameter></term>
<listitem>
<para>
Fetch all processes.
</varlistentry>
<varlistentry>
<term><parameter>GLIBTOP_KERN_PROC_PID</parameter></term>
<listitem>
<para>
Fetch process with pid <parameter>arg</parameter> or
nothing if no such process exists.
</varlistentry>
<varlistentry>
<term><parameter>GLIBTOP_KERN_PROC_PGRP</parameter></term>
<listitem>
<para>
Fetch only processes which are in process group
<parameter>arg</parameter>.
</varlistentry>
<varlistentry>
<term><parameter>GLIBTOP_KERN_PROC_SESSION</parameter></term>
<listitem>
<para>
Fetch only processes with session id
<parameter>arg</parameter>.
</varlistentry>
<varlistentry>
<term><parameter>GLIBTOP_KERN_PROC_TTY</parameter></term>
<listitem>
<para>
Fetch only processes with tty id
<parameter>arg</parameter>.
</varlistentry>
<varlistentry>
<term><parameter>GLIBTOP_KERN_PROC_UID</parameter></term>
<listitem>
<para>
Fetch only processes with effective user id
<parameter>arg</parameter>.
</varlistentry>
<varlistentry>
<term><parameter>GLIBTOP_KERN_PROC_RUID</parameter></term>
<listitem>
<para>
Fetch only processes with real user id
<parameter>arg</parameter>.
</varlistentry>
</variablelist>
<para>
You can use one or more of the following flags (with a
logical or) to hide some processes:
<variablelist>
<varlistentry>
<term><parameter>GLIBTOP_EXCLUDE_IDLE</parameter></term>
<listitem>
<para>
Don't fetch idle processes.
</varlistentry>
<varlistentry>
<term><parameter>GLIBTOP_EXCLUDE_SYSTEM</parameter></term>
<listitem>
<para>
Don't fetch system processes.
</varlistentry>
<varlistentry>
<term><parameter>GLIBTOP_EXCLUDE_NOTTY</parameter></term>
<listitem>
<para>
Don't fetch processes that have no controlling tty.
</varlistentry>
</variablelist>
</varlistentry>
<varlistentry>
<term><parameter>arg</parameter></term>
<listitem>
<para>
Additional argument depending upon the
<parameter>which</parameter> argument.
</varlistentry>
</variablelist>
<para>
This function returns a pointer (of type <type>unsigned *</type>) to
the list of process ids.
<para>
You have to manually &glibtop-free; this pointer once finished with it.
<!--
Local Variables:
mode: sgml
sgml-omittag:t
sgml-shorttag:t
sgml-minimize-attributes:nil
sgml-always-quote-attributes:t
sgml-indent-step:2
sgml-indent-data: t
sgml-parent-document: ("../gnome-hackers.sgml" "book" "sect1" "")
End:
-->

103
doc/features/procmem.sgml Normal file
View File

@@ -0,0 +1,103 @@
<para>
The automatically generated description above is taken from the
manual page of the <filename>/proc</filename> filesystem under Linux
and is a little bit confusing, so I make this clear here.
<note>
<title>Note for people porting &libgtop; to other systems</title>
<para>
Well, every operating system has its own idea about the memory usage
of a processes, and also system utilities like <filename>ps</filename>
show different things on different systems.
<para>
Nevertheless, we should try to make &libgtop; as system independent
as possible, so I give you some hints here how &glibtop-get-procmem;
should work.
<itemizedlist>
<listitem>
<para>
When you use &mmap; with either &MAP-SHARED; or &MAP-PRIVATE;,
this should only affect the &pmem-vsize; of the process and
none of its &pmem-size;, &pmem-resident;, &pmem-shared; and
&pmem-rss; sizes.
<listitem>
<para>
As soon as you read some of the &mmap;ed pages, they will be
demand-loaded and thus count towards the &pmem-size; of the
process. Also - we assume there is enough free memory - they
are resident in memory until they get stolen or swapped out
and thus increase the &pmem-resident; and &pmem-rss; sizes of
the process.
<listitem>
<para>
If the process has used &MAP-SHARED; and another process
attaches the same file also &MAP-SHARED; some of the pages
are shared with this process and thus increase the &pmem-shared;
sizes of both processes.
<listitem>
<para>
If the process has used &MAP-PRIVATE; and writes to the &mmap;ed
pages, the only difference to reading from them is that they
get dirty and cannot be stolen any longer but will get swapped
out.
<listitem>
<para>
If memory gets rare, clean pages are normally stolen which
decreases the &pmem-size;, &pmem-resident;, &pmem-shared; and
&pmem-rss; sizes of the process.
<listitem>
<para>
When dirty pages are swapped out, this will
<emphasis>not</emphasis> decrease the &pmem-size; of the
process but only its &pmem-resident; and &pmem-rss; sizes
(dirty pages cannot be shared).
<listitem>
<para>
The &pmem-vsize; of a process can <emphasis>only</emphasis>
be changed by the process itself when it requests or frees
memory but <emphasis>never</emphasis> due to swapping
activity of the system.
<listitem>
<para>
If the &pmem-shared; size changes, this
<emphasis>only</emphasis> means that the number of pages that
are currently shared with other processes has changed; if this
happens, this will <emphasis>never</emphasis> affect any of the
other sizes of the process.
</itemizedlist>
</note>
<para>
The hints above describe how it works under Linux - and we should try to
make &glibtop-get-procmem; show the same behavior under every other
system.
<para>
If you want to make any comments, flames, suggestions about this, please
feel free to do so.
<!--
Local Variables:
mode: sgml
sgml-omittag:t
sgml-shorttag:t
sgml-minimize-attributes:nil
sgml-always-quote-attributes:t
sgml-indent-step:2
sgml-indent-data: t
sgml-parent-document: ("../gnome-hackers.sgml" "book" "sect1" "")
End:
-->

23
doc/features/uptime.sgml Normal file
View File

@@ -0,0 +1,23 @@
<note>
<para>
You can calculate this from the CPU usage (this one is taken
from the FreeBSD port):
<programlisting>&include-uptime.txt;</programlisting>
</note>
<!--
Local Variables:
mode: sgml
sgml-omittag:t
sgml-shorttag:t
sgml-minimize-attributes:nil
sgml-always-quote-attributes:t
sgml-indent-step:2
sgml-indent-data: t
sgml-parent-document: ("../gnome-hackers.sgml" "book" "sect1" "")
End:
-->

40
doc/features/uptime.txt Normal file
View File

@@ -0,0 +1,40 @@
#include &lt;glibtop.h&gt;
#include &lt;glibtop/error.h&gt;
#include &lt;glibtop/uptime.h&gt;
#include &lt;glibtop/cpu.h&gt;
#include &lt;glibtop_suid.h&gt;
static const unsigned long _glibtop_sysdeps_uptime =
(1 << GLIBTOP_UPTIME_UPTIME) + (1 << GLIBTOP_UPTIME_IDLETIME);
static const unsigned long _required_cpu_flags =
(1 << GLIBTOP_CPU_TOTAL) + (1 << GLIBTOP_CPU_IDLE) +
(1 << GLIBTOP_CPU_FREQUENCY);
void
glibtop_get_uptime_p (glibtop *server, glibtop_uptime *buf)
{
glibtop_cpu cpu;
glibtop_init_p (server, GLIBTOP_SYSDEPS_UPTIME, 0);
memset (buf, 0, sizeof (glibtop_uptime));
/* We simply calculate it from the CPU usage. */
glibtop_get_cpu_p (server, &amp;cpu);
/* Make sure all required fields are present. */
if ((cpu.flags &amp; _required_cpu_flags) != _required_cpu_flags)
return;
/* Calculate values. */
buf->uptime = (double) cpu.total / (double) cpu.frequency;
buf->idletime = (double) cpu.idle / (double) cpu.frequency;
buf->flags = _glibtop_sysdeps_uptime;
}