Linux Power Management IEGD Considerations

Linux Power Management IEGD Considerations
White Paper
Mudit Vats
Graphic Software
Engineer
Intel Corporation
Ishu Verma
Software Technical
Marketing Engineer
Intel Corporation.
445352
Linux Power
Management
IEGD Considerations
March, 2010
Linux Power Management – IEGD Considerations
Executive Summary
Implementing Power Management (PM) is a complex undertaking
involving seamless interaction among several hardware and software
components. PM in Linux is even trickier as the open source choice
enables permutation of various PM players and solutions for a given
system.
Troubleshooting display-related issues when resuming from S3
suspend state in Linux is an extremely complex endeavor, requiring
an understanding of the handoff among various hardware and
software components.
When a system resumes from a power state, such as S3, handoffs among
several variables (BIOS, VBIOS, kernel, drivers, windows manager and X)
put the system back to its previous state. This white paper focuses on the
blank display (or no display) issue related to power state S3
suspend/resume.
With no display after resuming from S3 suspend state, the graphics driver
may seem to be the culprit. However, this behavior is often the result of
bugs in device drivers, BIOS, or VBIOS.
This paper presents some of the common techniques that you can use to
discover the root cause of the blank display issue, among them booting
headless, using an external graphics card, and using an open Linux
graphics driver. Some possible solutions involve disabling video repost
and optimizing xorg.conf.
The Intel® Embedded Design Center provides qualified developers with
web-based access to technical resources. Access Intel Confidential design
materials, step-by step guidance, application reference solutions, training,
Intel’s tool loaner program, and connect with an e-help desk and the
embedded community. Design Fast. Design Smart. Get started today.
www.intel.com/embedded/edc. §
2
Linux Power Management – IEGD Considerations
Contents
Power Management and Linux ..................................................................................4
Power States ........................................................................................4
S States (System States) .............................................................4
D States (Device States) ..............................................................5
Software Technologies ...........................................................................5
PM Framework ............................................................................5
X Windows System ......................................................................6
Window Managers .......................................................................6
PM Implementation with IEGD ..................................................................................7
Suspend to RAM ...................................................................................9
Resume/Wake-up ............................................................................... 11
What If Things Go Wrong? ..................................................................................... 12
System BIOS Issues ............................................................................ 12
Video BIOS ........................................................................................ 12
Miscellaneous Issues ........................................................................... 13
Diagnosing PM Issues .......................................................................... 13
Conclusion ........................................................................................................... 15
Glossary .............................................................................................................. 16
References........................................................................................................... 17
3
Power Management and Linux
Power management (PM) is a key feature for most computing platforms. PM
provides various benefits: increasing battery life; lowering heat emission;
lowering carbon footprint; and increasing the longevity of devices such as
panels, hard disk drives, and internal components. Implementing power
management is a complex undertaking involving seamless interaction among
several hardware and software components.
Before implementing PM, it is important to understand the available hardware
support for the various devices, among them CPU, DRAM, Flash, hard drive,
radio, Bluetooth*/WiFi/3G, and display.
Some examples of hardware PM include:
• CPU: HLT, Stop Clock, SpeedStep®
• Display: Blanking, dimming, Energy Star, power-save mode.
• Graphics Controller: Powering down (completely off or into an intermediate
power state).
• Hard Drive/CD-ROM: Spin down
• Network/NIC: Wake on LAN
• USB: Device power state transitions (mouse, USB drives, speakers); Wake
on access (mouse movement, R/W access, input signal)
• BIOS: ACPI S3/S4 platform behaviors, video repost, CPU PM settings
(states enabled, CPU fan throttling), and platform settings (e.g., thermal
low/high watermarks, chassis fan throttling).
Power States
There are various power states at the component and system level to enable
power savings. The following tables describe S States and D States in Linux.
S States (System States)
“S” States are responsible for platform-wide power state transitions.
445352
Linux Power Management – IEGD Considerations
Table 1. System States
State
Description
S1
CPU stops executing, cache is flushed; power to CPU and RAM is
maintained. Devices that can power down are powered down. Older
machines tend to support S1 versus S3.
S2
A deeper sleep. CPU is powered off. Not commonly implemented.
S3
“Standby” in Linux. “Sleep” in other operating systems. Also known as
“Suspend to RAM”.
S4
“Hibernation” in Windows. Also known as “Suspend to disk”.
S5
Soft power off. The system is off, but some power is available for
wake-up events (e.g., wake-on-LAN).
D States (Device States)
“D” States are responsible for device-specific power states transitions.
Table 2. Device States
State
Description
D0
Device is fully on.
D1
Intermediate state defined by device.
D2
Intermediate state defined by device.
D3
Device is powered off.
Software Technologies
The following sections describe various Linux software technologies that
implement some part of power management.
PM Framework
Linux supports two implementations of firmware power management:
Advanced Power Management (APM) and Advanced Configuration and Power
Interface (ACPI). APM is an older power management technology focused on
basic system and OS power management, with much of the power
management policy in the BIOS. It requires an APM driver as the interface
between the BIOS and the Linux operating system. With this method, power
management events could pass between the BIOS and the OS. Devices can
be notified of these events and respond appropriately. APM is still supported
in the kernel, but most newer platforms support ACPI, the successor to APM.
5
Linux Power Management – IEGD Considerations
The Linux kernel has supported ACPI for a long time. ACPI provides greater
power management state support, platform configuration support, and allows
for platform independence and OS control over power management events.
In those ways, it differs from APM, where BIOS was mostly in charge of
power management policies. ACPI supports power management, but also
thermal (fans), buttons/lids (events), CPU “C” states, and power source
(battery, AC), etc.
X Windows System
The X Window System provides graphics capabilities to the operating system.
It provides an infrastructure which allows Window Managers to provide a
desktop, widgets/controls, application windows, and toolbars, etc. The X
Window System has existed for many years and has been ported to multiple
operating systems. In terms of power management, the X Windows System
provides support for both APM and ACPI. For APM, it supports several of the
APM states. There’s also support for User/System Standby, User/System
Suspend, User/System Resume, and “Critical Standby/Suspend/Resume.”
When a power management event occurs, the X-Server is specifically notified
of system power management events. So, in this case, the X-Server
implements what to do in the case of power management events.
In terms of ACPI support, the X Windows System supports ACPI graphics
messages through acpid, but not system-wide power management events.
Because the X Window System is a user process, if the kernel mode driver
handles the suspend/resume power management correctly, the user mode
process of the X Window System should work like any other application when
put into and out of a power-managed state. To help facilitate state
transitions, the Enter/LeaveVT functions can be used to save or restore a
graphics state when transitioning from graphics, to console, and then to
suspend. The reverse is true as well when going from resume to console, and
then back into X.
Window Managers
As mentioned previously, Window Managers provide the graphical user
interface with which a user interacts. Linux supports many different window
managers. Two of the more popular window managers are GNOME and KDE.
These window managers have quite a bit of development support and many
distributions validate their distributions with these windowing solutions.
GNOME supports power management through the GNOME Power
Management solution, which uses hardware abstraction layer (HAL) open
source platform management, built on DBUS (open source messaging
interface). KDE3 supports KPowersave, its own power management solution.
Both of these window managers attempt to accomplish the same thing:
deliver reliable power management of the OS platform. OS vendors integrate
other third-party, open source technologies and include them in the
distributions, including scripts to ensure the correct sequence of events
6
Linux Power Management – IEGD Considerations
happen to successfully go into and out of a power managed state. Depending
on the solution, windows managers also may include a set of APIs that can be
used to hook into a particular power management solution.
Aside from these power management solutions for Linux, there are many
other “home grown” power management implementations, new windows
managers implementing even newer solutions, and several old and new
technologies. For example SWUSP, TuxOnIce, /sys/power/state,
/proc/acpi/event, and others, implement some part of Linux Power
Management. Having many power management players and solutions for
Linux is good. The complexity of it all is that “there are many power
management players and solutions for Linux”.
PM Implementation with IEGD
PM software manages state transitions along with device drivers and
applications. Device drivers are responsible for saving device states before
putting them into their low power states and then restoring the device state
when the system becomes active. Generally, applications are not involved in
power management state transitions. A few specialized applications that deal
directly with some devices may want handle state transitions. This section
explains the sequence of steps the Intel® Embedded Graphics Driver (IEGD)
for Linux goes through during power management state transitions.
In Figure 1. Power Management Sequence Diagram, the sequence can be
read by first understanding the top level blocks and then considering the
events that occur over time from top to bottom. The blocks are described as:
• Button/Peripheral: Describes input from a physical device such as a button,
key-press or mouse movement.
• BIOS: System BIOS responsible for platform configuration.
• VBIOS: Video BIOS in charge of pre-OS graphics initialization.
• Kernel: The Linux Kernel.
• Kernel (AGPGART) Driver: IEGD IKM (Intel Kernel Module) responsible for
graphics device initialization and resource allocation.
• X-Server Driver: The X Windows System graphics driver for an Intel
graphics device.
• Window Manager: The Graphical User Interface.
7
Linux Power Management – IEGD Considerations
Figure 1. Power Management Sequence Diagram
Button/Peripheral
BIOS
VBIOS
Kernel
Kernel (AGPGART) Driver
X Server Driver
Window Manager
Suspend to RAM
Switch to Console
LeaveVT, set to console video mode
Return
Freeze User Processes
Suspend
pm_suspend; Set to D3
Done
Freeze processors
Save Wake Vector
Save(WakeVector)
Power
Resume
Video Repost
Done
Reinitialize Video
Restore Wake Vector
Start OS Resume
Unfreeze processors
Resume
pm_resume, Set to D0
Unfreeze User Processes
Switch to Graphics
EnterVT, restore graphics mode
Done
The following two sections correspond to the sequence diagram and describe
Suspend to RAM and Resume/Wake-up.
8
Linux Power Management – IEGD Considerations
Suspend to RAM
When the system suspends or resumes from a power state such as S3, there
is handoff among several variables (BIOS, VBIOS, kernel, drivers, windows
manager and X) to put the system back to its previous state.
Examine what happens when the user puts the system into S3 suspend
mode, i.e., Suspend to RAM.
Note: Each distribution may have its own way of putting the system into a suspend mode
using any previously mentioned power management technologies available in Linux.
The Suspend to RAM action starts when an “event” is triggered in the
platform telling the OS that a power management event has occurred. The
event can be in the form of a button press, such as holding down the power
(or sleep) button on a laptop to trigger a sleep action or it can be through
user interaction, such as clicking on a window manager option to put the
platform into a suspended state. Certainly, one can execute a script at the
terminal or have a timer setup to automatically put the system in a low
power state as well. In the example above, assume the user chose a window
manager option to signal the GUI to change to a lower power state: Suspend
to RAM.
When this action occurs, the window manager tells the kernel to trigger an S3
power management operation. How this happens is distribution- and
implementation-specific. Newer Linux distributions have a platform
management infrastructure that provides a protocol for the OS to
communicate system or platform events between the software components
and the kernel. By using such an infrastructure, power management events
can be broadcast to other parts of the system through the kernel. Figure 1
shows that the window manager, after detecting the user-generated event,
triggers the kernel to go to a lower power state. The kernel then starts the
suspend to RAM procedure.
Alternatively, the suspend action triggers a hardware interrupt that is handled
by the embedded controller. The embedded controller triggers a register in
the southbridge chipset, setting a flag that a general purpose event has just
occurred. The OS notices this, and checks the DSDT for what must happen
next. Normally, this just calls a notification event that returns to user space
via /proc/acpi/events; the user space software determines what happens
next. Finally, the string "mem" is written to /sys/power/state.
The X display must switch to console before going into a lower power state.
With ACPI, the Linux kernel first switches to console mode and then continues
with its power management code. This is in contrast to APM where
suspend/resume often occurs through messages that were trapped from
within the X-Server. When X switches to the console, it calls a LeaveVT
(“leave virtual terminal”) function that allows the switch to occur. This is
where IEGD saves the graphics states.
9
Linux Power Management – IEGD Considerations
The process saves all graphics registers and state information so that on
resume, it can restore the graphics mode and state. After saving all of the
registers, the process restores the previous mode−the mode before the OS
entered into the GUI, i.e., console/terminal display mode. After restoring the
previous mode, the end-user sees a text mode display on the screen and the
kernel/OS can display text, which helps when debugging power management
issues.
After restoring the display to the console, the Linux kernel freezes all user
processes. This includes the X process and any other application running on
the system (window manager, browsers, X applications, etc.). With the user
applications frozen, the kernel can work with each device to ensure those
devices support the power management operation correctly and then invoke
the Suspend to RAM operation. After the user processes freeze, the kernel
probes each device driver and calls its associated suspend function (if
implemented). This gives each driver the opportunity to flush queues, stop
interrupts, process any remaining interrupts, and lower the power state of
the device. Some devices may or may not support putting the device into any
other lower power state other than D3 (powered off). Other devices may
support intermediate states, such as D1 or D2, i.e., the resulting power state
depends on the device capabilities.
In the case of IEGD, the driver puts the device into D3 no matter what kind
of power management event occurs. However, the driver sets D3 behavior in
the graphics controller by turning off the graphics device’s clocks on D3. On a
Suspend to RAM that makes a device operate in a lower power state.
After all devices have been put into a lower power state, the only code left
running is the kernel. The kernel does its own clean-up and freezes all
processors except for the one currently running. After running the kernel-side
suspend code, a couple of ACPI methods are executed: PTS (Prepare To
Sleep) and GTS (Going To Sleep). These executions tend to poke various
things that the kernel knows nothing about, and so a certain amount of
“magic” may be involved. At this point, the system should be dormant. Next,
the address of the kernel wakeup code is written to a location in the Fixed
ACPI Description Table (FADT). Finally, DSDT values are written to registers
described in the FADT. This usually causes some sort of system management
trap to ensure that the memory starts in self-refresh mode; it also sequences
the machine into suspend. For the S3 power state, this involves shutting the
machine down completely except for the RAM.
With the wake vector set, and user processes, devices drivers, and kernel all
in a “cleaned-up” state, the BIOS can now put the platform into a lower
power state.
Note: There is usually a setting in the BIOS that indicates what action to take for a suspend
operation. The options usually vary from a more aggressive power-managed state to
less aggressive power-managed state. In that case the OS instructs the BIOS that a
lower power state is desired. The BIOS often contains the policy for a lower power
state.
10
Linux Power Management – IEGD Considerations
Resume/Wake-up
When the user takes an action to wake the system, for example, a keystroke,
mouse movement, or a power button press, the system switches on. It then
jumps to the BIOS start address, does some setup such as programming the
memory controller, and then looks at the ACPI status register.
Note: Some types of Video BIOS support the ability of “video repost.” This allows the Video
BIOS to be called by the BIOS at the time of a resume operation (providing the BIOS
has support for this feature). Video repost is the full re-execution of the video BIOS
code. This helps when the graphics device is not working properly at the time of
resume.
The ACPI status register indicates that the machine was previously suspended
to RAM, so it then jumps to the wakeup address programmed earlier. The
wakeup address leads to real-mode x86 code provided by the kernel, which
programs the CPU back into protected mode and restores register state. At
this point, the kernel code continues execution.
From this point, the process is essentially the reverse of the suspend process.
The ACPI WAK method is called, all of the drivers are resumed, and user
space is restarted.
If X is running, Linux switches from console to the GUI. In the case of IEGD
and X, like the LeaveVT function, there is the analogous EnterVT (“enter
Virtual Terminal”), which restores the previously saved graphics mode−the
GUI display settings, saved register, and graphics device state information.
After EnterVT restores everything fully, it saves the previous mode, i.e., the
console mode before reentering the GUI.
With the kernel running again, all devices running, and user space
applications, including X, running again the system has successfully
resumed/woken-up from S3 Suspend to RAM.
11
Linux Power Management – IEGD Considerations
What If Things Go Wrong?
Considering the number of components involved in power management, there
are certainly many opportunities for failure. This section discusses common
areas of consideration and resolution techniques. Although many of the items
described are general to any Linux kernel, distribution, and devices in the
system, the focus is on the graphics device.
System BIOS Issues
BIOS is responsible for platform initialization and configuration. BIOS issues
can manifest themselves in a few ways. Some areas to consider are the
following:
• Bugs in the APIC code prevent restoring a single register, resulting in some
machines resuming without any interrupts being delivered.
• BIOS not setting a system register correctly or a register is being reset on
resume to some default and not being set again either by BIOS or kernel or
drivers. This can cause unpredictable results.
• Devices not being woken-up. On resume, power wells are not restored
correctly making some devices inaccessible. In these cases, the device has
power, but the device is no longer available to the system.
Video BIOS
Video BIOS (VBIOS)is a component of the platform’s firmware, which also
includes the System BIOS. Video BIOS is responsible for initializing and
configuring the graphics device. Typically, the VBIOS is the first option ROM
that executes in the platform firmware. This is essential, otherwise, no
display would be available at boot. Some VBIOS issues related to S3 resume
include the following:
• The ACPI specification does not require the BIOS to reprogram the video
hardware so it may come back in an entirely unprogrammed state.
Executing the code from VBIOS the same way the system BIOS does on
machine startup is a possible workaround.
• Typically video repost is necessary to help Linux bring the console back
after an S3; i.e., BIOS tells VBIOS to run its initialization code: video
repost. After VBIOS re-initializes, the console is back. With IEGD, this will
not work due to long standing and unchangeable architectural issues.
• In the case of VBIOS that supports video repost, e.g., GMA500 VBIOS, if it
is enabled AND IEGD’s IKM and X-Server are running, IEGD’s Graphics
Translation Tables (GTT) will be overrun. This results in incorrect mapping
of GTT entries and a corrupted display. The workaround is to disable video
repost in system BIOS.
12
Linux Power Management – IEGD Considerations
• Cache values have been issues with power management. For greater
performance, devices, such as IEGD tend to modify the default cache
values within the system, i.e., cache values for regions or memory to
achieve higher performance (see Wiki on MTRR for details). If the cache
values are reset on resume and are not restored correctly, the graphics
device performance may severely degrade, causing unusable, slow
conditions, or visual side-effects, or both.
Miscellaneous Issues
Some general OS issues include:
• A driver’s resume function may crash or cause a side-effect with another
driver. Resume functions are called in series, just like when devices are
suspended. A hang, crash, or any other issue can cause resource conflicts
for other devices. Sometimes this situation happens because PCI resources
can change at resume time.
• The Linux power management solution is inconsistent across various
hardware platforms. As a result, the power management for some platform
configurations work flawlessly and not for other configurations. This is an
area where OS vendors or system integrators provide value-added services
for integration and validation, including power management.
• Linux does not have a standard power management solution; however,
ACPI is common, and, to a lesser extent, APM is present. These days HAL
exists in some distros to support platform configuration, while others have
their own solutions, such as acpid, GNOME DBUS/HAL, KDE KPowerSave,
pm-utils for pm-suspend/pm-hibernate, SWUSP/TuxOnIce (S4), plus many
user-developed power management scripts. This significantly complicates
Linux power management.
Diagnosing PM Issues
It is often difficult to diagnose a power management issue because the
display may be a blank. Below are some methods that can help determine the
cause of power management problems.
• Disable the Intel graphics device. Enable SSH or VNC; i.e., boot headless.
Try suspend and resume remotely. If suspend and resume work without
issues, then there may be a graphics device issue, or a graphics driver
issue, or both.
• Disable the graphics device and try an external graphics device, such as
PCI or PCIE. Install the native driver and ensure that suspend and resume
work. The objective is to isolate Intel graphics device and driver issues. If
suspend/resume work fine, that may indicate an issue with the Intel
graphics device issue, or a graphics driver issue, or both.
13
Linux Power Management – IEGD Considerations
• Uninstall the IEGD driver. Use the non-embedded Intel GMA driver.
Observe suspend/resume results. If suspend/resume work fine, this implies
that the graphics device is working as it should, but there may be an IEGD
driver issue.
• Recompile the kernel for specific CPU/hardware, setting the correct options
for the processor. Take care to choose the right setting for the CPU that
correctly implements power management/speed throttling features. At this
stage, take this opportunity to fine-tune the kernel to choose the correct
options for the devices with the platform.
• If the VBIOS supports it and a BIOS option is available, enable video
repost. Try suspend and resume. Observe the state of the graphics device
after resume.
• If you are using the IEGD driver, including both IKM and graphics driver,
on an unsupported Linux distribution, ensure the IKM is correctly ported to
the kernel. Please see the white paper, IEGD Linux Kernel Module Patching and Porting Methods on
http://edc.intel.com/Software/Downloads/IEGD/#download for guidance.
• Tweak xorg.conf. IEGD supports a large number of graphics options.
Those options are configurable inside xorg.conf. The strategy: turn off any
non-critical features and create an xorg.conf with minimal configuration
for a single display with minimal features available. For debugging
purposes, turn off twin/clone support, hardware acceleration or any other
acceleration, and disable 2D acceleration and DRI (support for OpenGL),
etc. Disabling various features used by the IEGD driver decreases the
driver’s exposure to more complex areas of the graphic hardware.
• Try using the VGA or VESA graphics driver. To do this, modify xorg.conf to
use the standard VGA or VESA driver instead of “iegd” and then try
suspend/resume. If it works, that may provide evidence of the platform’s
ability to successfully go into and out of a power managed state.
• The IEGD driver supports several port types, such as LVDS, SDVO, Analog,
and HDMI. Because the IEGD driver uses different port drivers for each port
type, it may be valuable to try a different port types. This can help identify
port issues and potential display timing issues.
• The IEGD driver is highly customizable. The CED tool lets you configure
IEGD using a number of options. One key configuration option is the ability
to specify the display panel type and timings per the manufacturer’s
settings. Setting the timing correctly optimizes configuration for the display
device. This helps ensure the display is set correctly at each stage of the
power transitions.
The suggestions above may help diagnose and resolve power management
issues. Frequently on S3 Resume, for example, the system attempts to
resume but the screen will be blank, leading to the assumption that the
graphics driver is at fault. Sometimes the graphics device or graphics driver is
to blame, but not always. The hope is that by employing some of the
techniques above, IEGD customers can better identify the cause of power
management problems.
14
Linux Power Management – IEGD Considerations
Conclusion
Implementing PM in Linux requires understanding the complex interaction
among several hardware and software components in a given system. When
a system resumes from a power state, such as S3, there is state transition
among several variables (BIOS, VBIOS, kernel, drivers, windows manager,
and X) to put the system back to its previous state.
When there is lack of display after resuming from S3 suspend state, the
graphics driver may seem to be the cause. However, this behavior is usually
the result of bugs in device drivers, BIOS, or VBIOS, or some combination.
Some common techniques can be used to discover the root cause of the
blank display: booting headless, using an external graphics card, using open
Linux graphics driver, etc. Some of the possible solutions involve disabling
video repost and optimizing xorg.conf.
§
The Intel® Embedded Design Center provides qualified developers with webbased access to technical resources. Access Intel Confidential design
materials, step-by step guidance, application reference solutions, training,
Intel’s tool loaner program, and connect with an e-help desk and the
embedded community. Design Fast. Design Smart. Get started today.
http://intel.com/embedded/edc.
15
Linux Power Management – IEGD Considerations
Glossary
ACPI
Advanced Configuration and Power Interface. Besides being
the power management specification, it provides the OS with a
list of devices on the system. It also provides information
about interrupt routing, e.g., if someone has just removed a
hot-pluggable device from the system.
acpid
The user-space daemon needed to make the Linux ACPI
support completely functional.
AGPGART
A kernel module for Linux that supports the extra data transfer
features of Accelerated Graphics Port (AGP) video cards via the
Graphics Address Remapping Table (GART).
APM
Advanced Power Management.
DSDT
Differentiated System Description Table. The DSDT contains
the Differentiated Definition Block, which supplies the
implementation and configuration information about the base
system. DSDT is provided by system OEM. The OS always
inserts the DSDT information into the ACPI Namespace at
system boot time and never removes it.
The DSDT is in a bytecode called ACPI Machine Language
(AML), compiled from a simple language called ACPI Source
Language (ASL). At boot time, the system reads the DSDT,
parses it, and executes various methods. These can do pretty
much anything, but on the bright side they are being executed
in kernel context and, in principle, you can filter out anything
that you really do not want to do, such as overwriting CMOS.
The final relevant piece of ACPI information is something called
the FADT, or Fixed Address Descriptor Table. This gives the OS
information about various register addresses. It's a static
structure, and does not contain any executable code.
16
Embedded
Controller
The embedded controller performs complex low-level functions
through a simple interface to the host microprocessor.
FADT
Fixed ACPI Description Table (ACPI).
GNOME
A desktop environment that runs on top of the Linux kernel.
GTT
Graphics Translation Table.
HAL
A Hardware Abstraction Layer (HAL) is implemented in
software between the physical hardware of a computer and the
software running on that computer.
IEGD
Intel® Embedded Graphics Drivers.
Linux Power Management – IEGD Considerations
KDE
A free software project based around a cross-platform desktop
environment designed to run on Linux.
VBIOS
Video BIOS.
References
Matthew Garrett blog, “How Linux suspend and resume works in the ACPI
age,” posted February 7, 2007. http://www.advogato.org/article/913.html
Vaddagiri, Srivatsa, Anand K Santhanam, Vijay Sukthankar and Murali Iyer.
2004. Power Management in Linux Based Systems Linux Journal, no. 119
(March 2004).
ACPI Spec 3.0b – acpi.info
Authors
Mudit Vats is a Graphics Software Engineer at Intel Corporation.
Ishu Verma is a Software Technical Marketing Engineer at Intel
Corporation.
17
Linux Power Management – IEGD Considerations
INFORMATION IN THIS DOCUMENT IS PROVIDED IN CONNECTION WITH INTEL® PRODUCTS.
NO LICENSE, EXPRESS OR IMPLIED, BY ESTOPPEL OR OTHERWISE, TO ANY INTELLECTUAL
PROPERTY RIGHTS IS GRANTED BY THIS DOCUMENT. EXCEPT AS PROVIDED IN INTEL’S TERMS
AND CONDITIONS OF SALE FOR SUCH PRODUCTS, INTEL ASSUMES NO LIABILITY
WHATSOEVER, AND INTEL DISCLAIMS ANY EXPRESS OR IMPLIED WARRANTY, RELATING TO
SALE AND/OR USE OF INTEL PRODUCTS INCLUDING LIABILITY OR WARRANTIES RELATING TO
FITNESS FOR A PARTICULAR PURPOSE, MERCHANTABILITY, OR INFRINGEMENT OF ANY PATENT,
COPYRIGHT OR OTHER INTELLECTUAL PROPERTY RIGHT. Intel products are not intended for use
in medical, life saving, or life sustaining applications.
Intel may make changes to specifications and product descriptions at any time, without notice.
This paper is for informational purposes only. THIS DOCUMENT IS PROVIDED "AS IS" WITH NO
WARRANTIES WHATSOEVER, INCLUDING ANY WARRANTY OF MERCHANTABILITY,
NONINFRINGEMENT, FITNESS FOR ANY PARTICULAR PURPOSE, OR ANY WARRANTY OTHERWISE
ARISING OUT OF ANY PROPOSAL, SPECIFICATION OR SAMPLE. Intel disclaims all liability, including
liability for infringement of any proprietary rights, relating to use of information in this specification.
No license, express or implied, by estoppel or otherwise, to any intellectual property rights is granted
herein.
BunnyPeople, Celeron, Celeron Inside, Centrino, Centrino logo, Core Inside, Dialogic, FlashFile, i960,
InstantIP, Intel, Intel logo, Intel386, Intel486, Intel740, IntelDX2, IntelDX4, IntelSX2, Intel Core,
Intel Inside, Intel Inside logo, Intel. Leap ahead., Intel. Leap ahead. logo, Intel NetBurst, Intel
NetMerge, Intel NetStructure, Intel SingleDriver, Intel SpeedStep, Intel EP80579 Integrated
Processor, Intel StrataFlash, Intel Viiv, Intel vPro, Intel XScale, IPLink, Itanium, Itanium Inside, MCS,
MMX, Oplus, OverDrive, PDCharm, Pentium, Pentium Inside, skoool, Sound Mark, The Journey Inside,
VTune, Xeon, and Xeon Inside are trademarks or registered trademarks of Intel Corporation or its
subsidiaries in the U.S. and other countries.
*Other names and brands may be claimed as the property of others.
Copyright © 2010 Intel Corporation. All rights reserved.
18
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement