[PATCH 1/4] cpupower: Add cpupower-idle-info manpage

[PATCH 1/4] cpupower: Add cpupower-idle-info manpage

[Thomas Renninger]

The last missing manpage for cpupower tools.

More info about other architecture's sleep state specialities would be great.

Signed-off-by: Thomas Renninger <trenn@suse.de>
CC: linux-arm-kernel@lists.infradead.org
CC: jean.pihet@newoldbits.com
CC: linux-omap@vger.kernel.org
CC: linux-pm@lists.linux-foundation.org
CC: linux-acpi@vger.kernel.org
CC: cpufreq@vger.kernel.org
CC: linux@dominikbrodowski.net
---
 tools/power/cpupower/man/cpupower-idle-info.1 |   90 +++++++++++++++++++++++++
 1 files changed, 90 insertions(+), 0 deletions(-)
 create mode 100644 tools/power/cpupower/man/cpupower-idle-info.1

diff --git a/tools/power/cpupower/man/cpupower-idle-info.1 b/tools/power/cpupower/man/cpupower-idle-info.1
new file mode 100644
index 0000000..4178eff
--- /dev/null
+++ b/tools/power/cpupower/man/cpupower-idle-info.1
@@ -0,0 +1,90 @@
+.TH "CPUPOWER-IDLE-INFO" "1" "0.1" "" "cpupower Manual"
+.SH "NAME"
+.LP
+cpupower idle\-info \- Utility to retrieve cpu idle kernel information
+.SH "SYNTAX"
+.LP
+cpupower [ \-c cpulist ] idle\-info [\fIoptions\fP]
+.SH "DESCRIPTION"
+.LP
+A tool which prints out per cpu idle information helpful to developers and interested users.
+.SH "OPTIONS"
+.LP
+.TP
+\fB\-f\fR \fB\-\-silent\fR
+Only print a summary of all available C-states in the system.
+.TP
+\fB\-e\fR \fB\-\-proc\fR
+deprecated.
+Prints out idle information in old /proc/acpi/processor/*/power format. This
+interface has been removed from the kernel for quite some time, do not let
+further code depend on this option, best do not use it.
+
+.SH IDLE\-INFO DESCRIPTIONS
+CPU sleep state statistics and descriptions are retrieved from sysfs files,
+exported by the cpuidle kernel subsystem. The kernel only updates these
+statistics when it enters or leaves an idle state, therefore on a very idle or
+a very busy system, these statistics may not be accurate. They still provide a
+good overview about the usage and availability of processor sleep states on
+the platform.
+
+Be aware that the sleep states as exported by the hardware or BIOS and used by
+the Linux kernel may not exactly reflect the capabilities of the
+processor. This often is the case on the X86 architecture when the acpi_idle
+driver is used. It is also possible that the hardware overrules the kernel
+requests, due to internal activity monitors or other reasons.
+On recent X86 platforms it is often possible to read out hardware registers
+which monitor the duration of sleep states the processor resided in. The
+cpupower monitor tool (cpupower\-monitor(1)) can be used to show real sleep
+state residencies. Please refer to the architecture specific description
+section below.
+
+.SH IDLE\-INFO ARCHITECTURE SPECIFIC DESCRIPTIONS
+.SS "X86"
+POLL idle state
+
+If cpuidle is active, X86 platforms have one special idle state.
+The POLL idle state is not a real idle state, it does not save any
+power. Instead, a busy\-loop is executed doing nothing for a short period of
+time. This state is used if the kernel knows that work has to be processed
+very soon and entering any real hardware idle state may result in a slight
+performance penalty.
+
+There exist two different cpuidle drivers on the X86 architecture platform:
+
+"acpi_idle" cpuidle driver
+
+The acpi_idle cpuidle driver retrieves available sleep states (C\-states) from
+the ACPI BIOS tables (from the _CST ACPI function on recent platforms or from
+the FADT BIOS table on older ones).
+The C1 state is not retrieved from ACPI tables. If the C1 state is entered,
+the kernel will call the hlt instruction (or mwait on Intel).
+
+"intel_idle" cpuidle driver
+
+In kernel 2.6.36 the intel_idle driver was introduced.
+It only serves recent Intel CPUs (Nehalem, Westmere, Sandybridge, Atoms or
+newer). On older Intel CPUs the acpi_idle driver is still used (if the BIOS
+provides C\-state ACPI tables).
+The intel_idle driver knows the sleep state capabilities of the processor and
+ignores ACPI BIOS exported processor sleep states tables.
+
+.SH "REMARKS"
+.LP
+By default only values of core zero are displayed. How to display settings of
+other cores is described in the cpupower(1) manpage in the \-\-cpu option
+section.
+.SH REFERENCES
+http://www.acpi.info/spec.htm
+.SH "FILES"
+.nf
+\fI/sys/devices/system/cpu/cpu*/cpuidle/state*\fP
+\fI/sys/devices/system/cpu/cpuidle/*\fP
+.fi
+.SH "AUTHORS"
+.nf
+Thomas Renninger <trenn@suse.de>
+.fi
+.SH "SEE ALSO"
+.LP
+cpupower(1), cpupower\-monitor(1), cpupower\-info(1), cpupower\-set(1)
-- 
1.7.6.1

Re: [PATCH 1/4] cpupower: Add cpupower-idle-info manpage

[Jean Pihet]

Hi Thomas,

On Fri, Dec 16, 2011 at 3:35 PM, Thomas Renninger <trenn@suse.de> wrote:
> The last missing manpage for cpupower tools.
Great!

> More info about other architecture's sleep state specialities would be great.
I wonder if it is the right place for some arch specific information
about the sleep states.

I propose to document this in a more generic doc file (e.g.
Documentation/arm/OMAP/omap_pm) and link it from the manpage.

What do you think?

>
> Signed-off-by: Thomas Renninger <trenn@suse.de>
> CC: linux-arm-kernel@lists.infradead.org
> CC: jean.pihet@newoldbits.com
> CC: linux-omap@vger.kernel.org
> CC: linux-pm@lists.linux-foundation.org
> CC: linux-acpi@vger.kernel.org
> CC: cpufreq@vger.kernel.org
> CC: linux@dominikbrodowski.net

Regards,
Jean

Re: [PATCH 1/4] cpupower: Add cpupower-idle-info manpage

[Thomas Renninger]

On Friday 16 December 2011 17:18:37 Jean Pihet wrote:
> Hi Thomas,
> 
> On Fri, Dec 16, 2011 at 3:35 PM, Thomas Renninger <trenn@suse.de> wrote:
> > The last missing manpage for cpupower tools.
> Great!
> 
> > More info about other architecture's sleep state specialities would be great.
> I wonder if it is the right place for some arch specific information
> about the sleep states.
> 
> I propose to document this in a more generic doc file (e.g.
> Documentation/arm/OMAP/omap_pm) and link it from the manpage.
Duplicating info at different places increases needed maintenance.
Mabye above should be more about the kernel implementation details
and the manpage should more include the bits a typical user
wants to know?

> What do you think?
The POLL state is a bit confusing on X86 cpupower idle-info output.
Also on X86 the CPU may decide to use other states than the kernel
requested and the cpupower monitor tool should get used to see them.
This directly affects the output/usage of the tool and to understand it,
it should be documented in the manpage.

IMO it's not bad to have a somewhat bigger manpage here.
Especially on ARM, also the "ordinary" user wants to know
about how/whether the cpu power functionalities work as expected
without the need of looking at the kernel sources.

For now I could only try out cpupower on an ARM platform
with cpufreq support.
Would be great if someone could give cpupower idle-info
a try on such a machine.
Think of a typical user, not a kernel hacker and what he needs to
know to interpret and understand the output.
Maybe it's self-explanatory already, maybe the one or other arch
specific bit should get explained?

   Thomas