U.S. patent application number 13/865119 was filed with the patent office on 2013-09-05 for operating system context switching.
This patent application is currently assigned to Acer Cloud Technology, Inc.. The applicant listed for this patent is ACER CLOUD TECHNOLOGY, INC.. Invention is credited to Sandra Berndt, Miao Cui, Nigel Gamble, Wilson Ho, John Princen.
Application Number | 20130232327 13/865119 |
Document ID | / |
Family ID | 44859243 |
Filed Date | 2013-09-05 |
United States Patent
Application |
20130232327 |
Kind Code |
A1 |
Princen; John ; et
al. |
September 5, 2013 |
Operating System Context Switching
Abstract
A technique for quickly switching between a first operating
system (OS) and a second OS involves deactivating the first OS and
booting the second OS from memory. The technique can include
inserting a context switching layer between the first OS and a
hardware layer to facilitate context switching. It may be desirable
to allocate memory for the second OS and preserve state of the
first OS before deactivating the first OS and booting the second OS
from memory.
Inventors: |
Princen; John; (Cupertino,
CA) ; Berndt; Sandra; (Cupertino, CA) ; Cui;
Miao; (Los Altos, CA) ; Gamble; Nigel;
(Mountain View, CA) ; Ho; Wilson; (San Mateo,
CA) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
ACER CLOUD TECHNOLOGY, INC. |
Sunnyvale |
CA |
US |
|
|
Assignee: |
Acer Cloud Technology, Inc.
Sunnyvale
CA
|
Family ID: |
44859243 |
Appl. No.: |
13/865119 |
Filed: |
April 17, 2013 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
12784382 |
May 20, 2010 |
8433889 |
|
|
13865119 |
|
|
|
|
61329066 |
Apr 28, 2010 |
|
|
|
Current U.S.
Class: |
713/2 |
Current CPC
Class: |
G06F 9/461 20130101;
Y02D 10/00 20180101; Y02D 10/28 20180101; Y02D 10/26 20180101; G06F
9/4406 20130101; G06F 9/45533 20130101 |
Class at
Publication: |
713/2 |
International
Class: |
G06F 9/44 20060101
G06F009/44 |
Claims
1. A system comprising: a context switching layer insertion engine;
an operating system (OS) memory allocation engine coupled to the
context switching layer insertion engine; a memory boot engine
including a processor, the memory boot engine coupled to the
context switching layer insertion engine; wherein, in operation:
the context switching layer insertion engine inserts a context
switching layer between a first OS and a hardware layer; the OS
memory allocation engine allocates memory for a second OS and loads
data associated with the second OS into the memory; the memory boot
engine boots the second OS from memory.
2. The system of claim 1, wherein the context switching layer
comprises a minimalist hypervisor.
3. The system of claim 1, wherein inserting the context switching
layer comprises intercepting a control signal from the first OS or
the second OS, and the hardware layer.
4. The system of claim 1, wherein inserting the context switching
layer comprises: receiving, at the context switching layer
insertion engine, a first signal from the first OS or the second
OS; providing, from the context switching layer insertion engine, a
second signal to the hardware layer, the second signal appearing to
the hardware layer as if the second signal were coming from the
first OS or the second OS.
5. The system of claim 4, wherein the first signal comprises a
request for hardware resources from the hardware layer.
6. The system of claim 1, wherein inserting the context switching
layer comprises: receiving an indication from the first OS that the
first OS is to go to sleep; intercepting a write request from the
first OS to the hardware layer; setting, in the hardware layer, a
sleep type or sleep enable bits associated with the first OS.
7. The system of claim 6, wherein the write request comprises one
or more requests to a power management block of the hardware
layer.
8. The system of claim 1, wherein the context switching layer
comprises an input/output (I/O) driver configured to enable saving
a state of the first OS to a datastore.
9. The system of claim 1, wherein, in operation: the context
switching layer insertion engine inserts another context switching
layer between the second OS and the hardware layer; the OS memory
allocation engine allocates another memory for the first OS and
loads data associated with the first OS into the other memory; the
memory boot engine boots the first OS from the other memory,
thereby restoring the first OS.
10. A method comprising: inserting a context switching layer
between a first OS and a hardware layer; allocating memory for a
second OS; loading data associated with the second OS into the
memory; booting the second OS from memory.
11. The method of claim 10, wherein the context switching layer
comprises a minimalist hypervisor.
12. The method of claim 10, wherein inserting the context switching
layer comprises intercepting a control signal from the first OS or
the second OS, and the hardware layer.
13. The method of claim 10, wherein inserting the context switching
layer comprises: receiving, at the context switching layer
insertion engine, a first signal from the first OS or the second
OS; providing, from the context switching layer insertion engine, a
second signal to the hardware layer, the second signal appearing to
the hardware layer as if the second signal were coming from the
first OS or the second OS.
14. The method of claim 13, wherein the first signal comprises a
request for hardware resources from the hardware layer.
15. The method of claim 10, wherein inserting the context switching
layer comprises: receiving an indication from the first OS that the
first OS is to go to sleep; intercepting a write request from the
first OS to the hardware layer; setting, in the hardware layer, a
sleep type or sleep enable bits associated with the first OS.
16. The method of claim 15, wherein the write request comprises one
or more requests to a power management block of the hardware
layer.
17. The method of claim 10, wherein the context switching layer
comprises an input/output (I/O) driver configured to enable saving
a state of the first OS to a datastore.
18. The method of claim 10, further comprising: inserting another
context switching layer between the second OS and the hardware
layer; allocating another memory for the first OS and loads data
associated with the first OS into the other memory; booting the
first OS from the other memory, thereby restoring the first OS.
19. A system comprising: at least one processor; memory storing
instructions configured to instruct the at least one processor to
perform: inserting a context switching layer between a first OS and
a hardware layer; allocating a portion of the memory for a second
OS; loading data associated with the second OS into the portion of
the memory; booting the second OS from the portion of the
memory.
20. The system of claim 19, wherein the context switching layer
comprises a minimalist hypervisor.
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application is a continuation of U.S. patent
application Ser. No. 12/784,382, filed May 20, 2010 and entitled
"Operating System Context Switching," now U.S. Pat. No. 8,433,889,
which claims priority to U.S. Provisional Patent Application Ser.
No 61/329,066, filed Apr. 28, 2010 and entitled "Operating System
Context Switching," which applications are incorporated by
reference herein.
BACKGROUND
[0002] A computer is a programmable machine that receives input,
stores and manipulates data, and provides output in a useful
format. Computers can be found in a huge number of devices that
could not realistically be exhaustively listed, including by way of
example personal computers (PCs), smart phones, game consoles, web
servers, personal data assistants (PDAs), automobiles, etc. Nearly
all modern computers implement some form of stored-program
architecture, which is the single trait by which computers are now
defined.
[0003] In the 1960s libraries of common programs were stored on
computers to increase efficiency when programs were run in
succession. Eventually, computers were able to automatically switch
between tasks, and runtime libraries were created to manage
processing and printing speed. These and other features are now
referred to as the operating system (OS). Common features for OSs
include process management, interrupts, memory management, virtual
file systems, device drivers, networking, security, and graphical
user interfaces.
[0004] Computers that are capable of running multiple programs
typically include an OS. The main purpose of the OS is to manage
the way the programs use the hardware, and regulates the way that a
user controls the computer. Some popular OSs for PCs include
Microsoft Windows, Mac OS X, and Linux, but there are a great many
others. The central component of most OSs is the kernel, which is a
bridge between programs and data processing done at the hardware
level. The kernel can provide the lowest-level abstraction layer
for computer resources, especially processors and I/O devices, that
programs need to perform their function.
[0005] It is possible to use multiple OSs on a single computer by
changing the boot loader to allow a user to select the operating
system they want to run at boot time. However rebooting is slow. It
is also possible to use multiple OSs on a single computer by using
virtualization to run multiple guest operating systems on top of a
single host operating system. However, virtualization does not
generally allow each OS direct access to hardware devices, which
can reduce performance, particularly for programs that require
direct access to high speed peripherals, such as games.
[0006] The foregoing examples of the related art and limitations
related therewith are intended to be illustrative and not
exclusive. Other limitations of the related art will become
apparent upon a reading of the specification and a study of the
drawings.
SUMMARY
[0007] The following examples and aspects thereof are described and
illustrated in conjunction with systems, tools, and methods that
are meant to be exemplary and illustrative, not limiting in scope.
In various examples, one or more of the above-described problems
have been reduced or eliminated, while other examples are directed
to other improvements.
[0008] A technique for quickly switching between a first operating
system (OS) and a second OS involves deactivating the first OS and
booting the second OS from memory. The technique can include
inserting a context switching layer between the first OS and a
hardware layer to facilitate context switching. The technique can
include allocating memory for the second OS, which can include
arranging memory contents to put the second OS in place. The
technique can include preserving state of the first OS if it is
desirable to switch back to the first OS at a later time.
[0009] In an embodiment, the context switching layer can include a
hypervisor. A hypervisor normally presents a guest OS with a
virtual platform and monitors the execution of the guest OS, which
enables multiple OSs to share hardware resources. Advantageously,
when the hypervisor is used in a context switching layer, the
hypervisor can facilitate switching from the first OS to the second
OS without necessarily requiring ongoing virtualization, which can
result in improved performance. Advantageously, booting the second
OS from memory can occur without rebooting through boot firmware,
such as the basic input/output system (BIOS), extensible firmware
interface (EFI), or other code that is run when a computer is
powered on.
[0010] These and other advantages will become apparent to those
skilled in the relevant art upon a reading of the following
descriptions and a study of the several examples of the
drawings.
BRIEF DESCRIPTION OF THE DRAWINGS
[0011] The following figures are intended to illustrate by way of
example some aspects of techniques described in this paper.
[0012] FIG. 1 depicts a conceptual diagram of operational layers of
a fast context switching system.
[0013] FIG. 2 depicts an example of a system capable of fast
context switching.
[0014] FIG. 3 depicts an example of an OS switching system.
[0015] FIG. 4 depicts a flowchart of an example of a method for
fast context switching.
[0016] FIG. 5 depicts an example of a computer system.
DETAILED DESCRIPTION
[0017] In the following description, several specific details are
presented to provide a thorough understanding. One skilled in the
relevant art will recognize, however, that the concepts and
techniques disclosed herein can be practiced without one or more of
the specific details, or in combination with other components, etc.
In other instances, well-known implementations or operations are
not shown or described in detail to avoid obscuring aspects of
various examples disclosed herein.
[0018] FIG. 1 depicts a conceptual diagram 100 of operational
layers of a fast context switching system. The diagram 100 includes
a hardware layer 102, a fast context switching layer 104, an
operating system (OS) layer 106, and an application layer 108.
[0019] In the example of FIG. 1, the hardware layer 102 is the
bottom layer, upon which other layers are built. The hardware layer
102 includes a processor and can include other hardware components,
such as additional processors and networking, graphics, and disk
hardware. Conceptually, it is not important whether the hardware on
which components of the other layers are implemented is part of the
hardware layer 102, and this may or may not be the case for any
given implementation. This is at least in part because, for
example, programs at the application layer 108 can be implemented
as software embodied in a computer-readable medium that includes
hardware in a manner that is well-understood to those of skill in
the relevant art. It should be noted that a person of skill in the
relevant art would have no difficulty understanding what is meant
by the term "computer-readable medium." To the extent legal
interpretation introduces ambiguity, it should be understood that
the intention in this paper is to encompass all statutory
computer-readable media and to exclude nonstatutory interpretations
of the term "computer-readable medium" to the extent it is
necessary to preserve claim validity. Further, if the law changes,
the term computer-readable medium should be interpreted consistent
with the law of the time. Since the law is presumably unambiguous
at any given time, such as today, the meaning of the term
"computer-readable medium" is both clear to one of skill in the
relevant art and not ambiguous to one with knowledge of the law
since the definition encompasses only computer-readable media that
is allowed under the law, no more and no less. "Computer readable
storage medium" can also be used to refer to known or convenient
statutory computer readable media, which are currently understood
to exclude non-tangible media.
[0020] In the example of FIG. 1, the fast context switching layer
104 is conceptually located on top of the hardware layer 102. The
other layers of the system 100 may or may not be agnostic with
respect to the fast context switching layer 104. For example, an OS
in the OS layer 106 may or may not be aware of the fast context
switching layer interposed between the OS and the hardware layer
102. Thus, advantageously, fast context switching can be introduced
into a system without changing applications, OS, or hardware. The
fast context switching layer 104 includes components that enable
fast switching between OSs using techniques that are described
later.
[0021] In the example of FIG. 1, the OS layer is conceptually
located on top of the fast context switching layer 104. Typically,
an OS is implemented as software embodied in a computer-readable
medium, which can include firmware, hardware, or a combination
thereof. The OS layer 106 includes an active OS 110 and inactive OS
112-1 to 112-N (collectively, the inactive OSs 112). The active OS
110 has at least occasional bare metal access to the hardware layer
102. This can be accomplished by requiring that the fast context
switching layer 104 be inserted in a minimalist fashion using
techniques that are described later. The OS layer can swap between
the active OS 110 and one of the inactive OSs 112 using fast
context switching techniques described later. Advantageously, none
of the OSs are treated as "guest" OSs that are forced to operate on
top of a virtualization layer, such as a "thick" hypervisor. As
used in this paper, "fast" context switching means that switching
between OSs does not require a full reboot using boot firmware,
such as BIOS, EFI, or other code. It should be noted that the OS
layer 106 at a given time actually comprises only the active OS
110, since the inactive OSs 112 are not interposed between the fast
context switching layer 104 and the application layer 108. This is
different from, for example, a system that includes multiple guest
OSs, which would be stacked on top of a master OS, and an
application in the application layer would, at least in some cases,
have to pass through the guest OS and the master OS to reach
hardware resources at the hardware level (of course, there would be
no fast context switching layer).
[0022] In the example of FIG. 1, the application layer 108 is
conceptually located on top of the OS layer 106. The OS layer 106
enables applications in the application layer 108 to make use of
components of the hardware layer 102 that it needs to carry out its
functionality. Typically, an application is implemented as software
embodied in a computer-readable medium. Typically, an application
can provide output and receive input through drivers in the OS
to/from hardware in the hardware layer.
[0023] FIG. 2 depicts an example of a system 200 capable of fast
context switching. The system 200 includes hardware layer 202, OS
204, application layer 206, context switching engine 208, OS data
store 210, and OS 212. The hardware layer 202, OS 204, and the
application layer 206 are similar to the hardware layer 102, active
OS 110, and application layer 108 described previously with
reference to FIG. 1.
[0024] In the example of FIG. 2, the context switching engine 208
is conceptually located above the hardware layer 202 and below the
OS 204, but does not necessarily lie between the hardware layer 202
and the OS 204 (at least not necessarily all of the time). As used
in this paper, an engine includes a processor and memory including
memory modules that can be executed by the processor. The context
switching engine 208 can be implemented as software embodied in a
computer-readable medium in the OS 204, partially in the OS 212 if
the OS 212 supports memory hot-plugging, a logical layer between
the hardware layer 202 and the OS 204, 212, or a combination of
these. For example, the context switching engine 208 can include a
driver in the kernel of the OS 204, triggered by a switch
instruction from a context-switching application. (The
context-switching application can be considered "part of" the
context switching engine 208 in the sense that actions taken by the
context-switching application can be described in this paper as
actions taken by the context switching engine 208.) The context
switching engine 208 can use virtualization features of a
processor, such as VMexit to exit into a hypervisor and VMenter to
enter a virtual machine. Alternatively, OS code could be rewritten
to call a new context without these virtualization features, though
this technique would still call for a logical separation of the OS
from the hardware. The context switching engine 208 can use power
management features of an OS, such as advanced configuration and
power interface (ACPI) sleep functionality. The Advanced
Configuration and Power Interface Specification Revision 4.0, Jun.
16, 2009, is incorporated by reference. In an implementation, the
context switching engine 208 can trigger the sleep function of an
OS, but intercept a final write to the OS. So the sleep
functionality can properly be characterized as part of the OS or
logically as part of the context switching engine 208 even though
some implementations may simply have the context switching engine
208 trigger the sleep functionality and intercept a final write,
but rely upon the OS to carry out the sleep. A final write can
include, for example, one or more writes to a power management
block, which could be implemented using an fixed ACPI description
table (FADT) power management 1 (PM1). (PM1 refers to power
management associated with sleep.) Normally, the final write would
power down hardware after stopping processes, I/O, and application
processors, leaving a single processor executing. (In the ACPI
example, the OS would set the sleep type and enable bits in the
FADT PM1 control register.) However, since the context switching
engine 208 intercepts the final write, the context switching engine
208 can instead trigger a boot loader. The context switching engine
208 can be referred to as creating or inserting a context switching
layer between the hardware layer 202 and the OS 204.
[0025] In the example of FIG. 2, the OS data store 210 is coupled
to the context switching engine 208. The OS data store 210 can be
implemented as one or more flat files, databases, and/or other data
structures embodied in a computer-readable medium. In operation,
the OS data store 210 includes data associated with the OS 204 and
the OS 212, including OS state. The data should include sufficient
information to enable the context switching engine 208 to restore
an OS to an initial or previous state. For example, system state
that is in memory can remain in memory (which would mean that the
OS data store 210 includes at least part of system memory) or it
can be stored to disk to save memory resources, or some combination
such as leaving the system state in memory until memory resources
are needed, at which point the system state could be saved to disk
and the memory locations reused. As another example, power state,
such as ACPI FADT/firmware ACPI control structure (FACS) state,
could be stored in registers. Since the context switching engine
208 switches from the OS 204 to the OS 212, it would be desirable
to save the power state elsewhere. Normally when a system goes to
sleep, the data in the registers does not have to be saved
elsewhere because it is assumed that when the system wakes up again
it will be in the same OS. So typical OS sleep functionality does
not include saving power state. However, when the system 200 "wakes
up," it will have switched from the OS 204 to the OS 212,
necessitating saving the data in the registers to some other
location due to the fact that the OS 212 will overwrite the values
in the registers. Depending upon the implementation, the OS data
store 210 could include data for each OS in a similar format, in
the same data structure, in different formats/data structures
depending upon whether the state is an initial state or a previous
state, whether the state is system state or power state, etc. When,
where, and how data is stored is not critical so long as the OS
data store 210 enables the context switching engine 208 to perform
its switching function.
[0026] In the example of FIG. 2, in operation, only one of the OS
204 and the OS 212 is active at a time. Therefore, advantageously,
neither the OS 204 nor the OS 212 are treated as a "guest" OS,
regardless of which OS is booted at startup. Moreover, there is no
need for a "thick" hypervisor or virtual machine monitor to
persistently sit between the OS 204, 212 and the hardware layer
202. Rather, the context switching engine 208 can insert a context
switching layer only as needed (presumably as late as possible to
maximize performance). Alternatively, the context switching engine
208 could create a context switching layer that is designed to
facilitate optimal performance by allowing the OS to run "on bare
metal" as much as possible. The context switching layer can be
implemented, for example, as a "minimalist" hypervisor that is
limited in time (with a goal of minimum amount of time) between the
OS and the hardware or limited in the amount of separation between
the OS and hardware layers (with a goal of minimum amount of
separation). For example, conceptually, the context switching
engine 208 virtualizes only those elements of the hardware that are
needed to accomplish context switching and/or only puts those
elements in place when they are needed.
[0027] In the example of FIG. 2, in operation, the context
switching engine 208 can receive a switch instruction as input. The
switch instruction tells the context switching engine 208 to switch
from the OS 204 to the OS 212. The switch instruction can be
provided in a number of different ways. For example, a user could
trigger the switch instruction by launching a context-switching
application that triggers the switch instruction, the OS 204 (or
the OS 212) could decide when to trigger the switch instruction, or
the switch instruction could be provided in response to some other
stimulus.
[0028] In the example of FIG. 2, in operation, the context
switching engine 208 inserts a context switching layer between the
hardware layer 202 and the OS 204. (The context switching layer is
considered "part of" the context switching engine 208 in the sense
that actions taken at the context switching layer can be described
in this paper as actions taken by the context switching engine
208.) To accomplish this task, the context switching engine 208 is
capable of allocating memory for the context switching layer and
loading it. For example, a context-switching application could
trigger a kernel driver to insert a context switching layer between
the hardware layer 202 and the OS 204, and set up the context
switching layer to catch accesses to the OS address space. In an
embodiment, the context switching layer is inserted in response to
the switch instruction. Alternatively, the context switching layer
could be inserted prior to receiving the switch instruction as a
thin but persistent layer.
[0029] In the example of FIG. 2, for illustrative purposes, the OS
204 is initially active when the context switching engine 208
receives the switch instruction. The context switching engine 208
allocates memory to load the OS 212 and relevant OS state, data for
which can be obtained from the OS data store 210. To accomplish
this task, the context switching engine 208 is capable of
allocating memory for a next OS (the OS 212 in this example) and
the state associated with the next OS, and loading it. The context
switching engine 208 can also arrange memory contents to put the OS
212 in a place where it can be booted. For example, the context
switching engine 208 (in an embodiment, a kernel driver) can
allocate memory for an OS image associated with the OS 212, and
load the image in preparation for boot. In an embodiment, the
memory is allocated in response to the switch instruction.
[0030] In the example of FIG. 2, in operation, the context
switching engine 208 puts the OS 204 to sleep. Sleep is usually
fast because it often does not involve storing state to disk. Sleep
under normal circumstances will shut down all drivers, and keep all
state in memory, power down most peripherals and then power down
the main CPU, while maintaining power to memory. To accomplish this
task, the context switching engine 208 can instruct the OS to sleep
which, for example, can cause the OS to freeze all processes, stop
device I/O, leave system state in memory, freeze application
processors, and leave only a single processor running; then write
to a power management block that is part of the hardware and cause
the system to power down. If the system 200 is ACPI-compliant, the
OS 204 may have a FADT PM1 to which the command is written. In this
example, the OS 204 then sets the sleep type and the sleep enable
bits in the FADT PM1 control register. Advantageously, the OS 204
need not know about the context switching and can use functionality
that is present in most modern OSs to put itself to sleep.
[0031] In the example of FIG. 2, in operation, the context
switching engine 208 saves the state of the OS 204 (and perhaps
other data associated with the OS 204) in the OS data store 210. It
is expected that the context switching engine 208 will save the
state in response to the switch instruction, though it is also
possible that state could be saved continuously or
occasionally.
[0032] In an embodiment, the context switching engine 208 tells the
OS 204 to go to sleep, the OS 204 will begin to put itself to
sleep, data associated with the OS 204, such as system state, is
saved, then the context switching engine 208 can catch instructions
right before the OS 204 goes to sleep and perform context switching
instead. To accomplish this task, the context switching engine 208
can catch a final write to the FADT PM1 control register at the
context switching layer, but instead of powering down the system,
trigger a boot loader.
[0033] In the example of FIG. 2, the context switching engine 208
sets up the machine for booting the OS 212. The context switching
engine 208 can, for example, trigger a boot loader to disable
interrupts, disable paging or remap memory, save the power state
for the OS 204, swap the contents in low memory (0 to N Mbytes) and
the OS 212 memory contents, restore or set the power state for the
OS 212, and setup the machine for booting the OS 212. (The boot
loader can be considered "part of" the context switching engine
208.) The power state can include, for example, ACPI FADT/FACS
state. The power state may represent the contents of registers that
are normally left in place when an OS goes to sleep, but which must
be saved elsewhere if the system switches to a different OS.
[0034] In the example of FIG. 2, in operation, the OS 212 is then
booted from memory. To accomplish this task, the context switching
engine 208 (specifically, the boot loader) can jump to the next OS
in memory, which is the OS 212 in this example, to start it. Start
up includes initializing hardware devices the OS 212 needs to
operate, and saving any state necessary from the previous OS, which
is the OS 204 in this example, so that the previous OS can be
restarted later, if applicable. Advantageously, using this
technique, the OS 212 can be booted without using the entire boot
firmware, such as BIOS, EFI, or other code. So context switching is
relatively fast; booting an OS without using the entire boot
firmware can be referred to as fast context switching.
[0035] In the example of FIG. 2, at this point in operation, the OS
212 is fully operational and can access the hardware layer 202
directly at full performance. State from the OS 204 may or may not
be maintained in memory. For example, it may be desirable to
implement a system that leaves OS 204 state in memory, unless the
memory resources need to be recycled. If the OS 212 needs more
memory and there are no unused areas, the OS 212 could write out
memory areas from the sleeping OS 204 to disk, and free up those
areas for use. Alternatively, it may be desirable to implement a
system in which the context switching engine stores system state to
the OS data store 210 and recycles memory resources as a matter of
course. As another example, power state might be saved to the OS
data store 210 by the context switching engine in case of switch
back to the OS 204.
[0036] To restore the OS 204, the process described above for
switching from the OS 204 to the OS 212 is reversed. For example,
the context switching engine 208 shuts down applications and
services, frees up pages, restores state from disk, inserts a
context switching layer under the OS 212, shuts down drivers
managing hardware; the context switching layer catches accesses to
a power management block that would normally cause power down; and
the context switching engine 208 restores memory areas that were
moved away to allow the OS 212 to run and executes the resume from
sleep entry point that will cause the OS 204 to run again. In this
way, the OS 204 and driver, application, and power state is
restored as before a context-switching application was launched to
trigger the switch from the OS 204 to the OS 212.
[0037] The system 200 is described as allowing fast context
switching midstream while the OS 204 is active. However, using this
technique, an alternative boot loader could be implemented to
change the boot firmware so as to avoid a full reboot and start
with any desired OS. In this way, the default OS can be any OS that
is designated by the alternative boot loader, and the system 200
can start up in the default OS. Thus, it is possible to accomplish
context switching that spans a restart (or spans power down to
startup) following an adjustment to the alternative boot loader in
a previous session.
[0038] FIG. 3 depicts an example of an OS switching system 300. The
system 300 includes a context switching layer insertion engine 302,
an OS memory allocation engine 304, an OS power-save trigger engine
306, an OS state preservation engine 308, an OS power-down
prevention engine 310, a memory boot engine 312, and an optional
memory recycling engine 314. In the example of FIG. 3, initially, a
first ("previous") OS is active, as illustrated by the box 315.
[0039] In the example of FIG. 3, the context switching layer
insertion engine 302 receives a switch instruction and creates a
context switching layer 316 for insertion between the first OS and
a hardware layer (not shown). By separating the hardware from the
OS, the context switching layer 316 can introduce additional
functionality while the OS and the hardware can operate without
recoding or redesign. The context switching layer can be used to
facilitate fast switching from one OS to another; so the context
switching layer can be minimalist in the sense that it is inserted
only when needed and/or only to the extent it is need to facilitate
switching of OSs. The context switching layer 316 can include, for
example, a minimalist hypervisor.
[0040] When data (including, e.g., control signals) pass from the
OS to the hardware or vice versa, the data passes through the
context switching layer. The context switching layer can also
generate signals "as if" coming from the OS or the hardware, and
can intercept signals to prevent them from passing from the OS to
the hardware or vice versa. For example, the context switching
layer could trigger an OS to go to sleep, allow the OS to carry out
the process, but intercept a final write from the OS to the
hardware (specifically, a control register) to, e.g., set sleep
type and/or sleep enable bits. The context switching layer can also
be given other functionality, such as an I/O driver to enable the
saving of state or other data to a storage device, though in at
least one implementation the context switching layer has no drivers
(in such an implementation, to the extent drivers are needed, they
are utilized through an OS).
[0041] In the example of FIG. 3, the OS memory allocation engine
304 allocates memory for a second ("next") OS, as illustrated by
the box 318. The OS memory allocation engine 304 loads data
associated with the second OS into memory. The memory allocation
and loading of data associated with the second OS does not, in and
of itself, cause the system to reboot with the second OS.
Typically, OS data is in low memory (currently occupied by the
first OS). If the second OS data was simply loaded into low memory
while the first OS was running, it would probably cause a system
crash, if it could be done at all. For this reason, the data
associated with the second OS is loaded into memory with the
intention of rearranging the memory contents such that the second
OS can take over from the first OS without causing a system crash
and without the need to shut down the system and reboot with the
second OS.
[0042] In the example of FIG. 3, the OS power-save trigger engine
306 causes the first OS to initiate power-save (e.g., sleep)
procedures, as illustrated by the box 320. Most modern OSs and
hardware enable entering a power-save mode, such as sleep. So the
OS power-save trigger engine 306 can potentially only need to
trigger sleep procedures that are already available through an OS.
If the OS did not have sleep capabilities, but was implemented on
hardware capable of entering sleep or another known or convenient
power-save or power-down state, the OS power-save trigger engine
306 could include additional functionality that would cause the
hardware to begin to power down. In this paper, a power-save mode
is intended to mean a sleep mode, which is one example of a
power-save mode, or any applicable known or convenient power-save
mode. Since power-save modes cause a system to reduce its power,
the system can be said to be "powering down" when it initiates
entry into a power-save mode, even if ultimately the system does
not power down completely.
[0043] In the example of FIG. 3, the OS state preservation engine
308 stores state information associated with the first OS in an OS
state data store 322. The OS state in the OS state data store 322
can include system state, power state, and other state associated
with inactive OSs. It may be the case that the OS state
preservation engine 308 stores different state at different times.
For example, the OS state preservation engine 308 might store
system state after the OS power-save trigger engine 306 causes the
first OS to begin a power-save procedure, but might store power
state after the OS power-down prevention engine 310 prevents the
system from completing entry into the power save mode.
[0044] In the example of FIG. 3, the OS power-down prevention
engine 310 prevents the first OS from completing entry into a power
save mode. The first OS will be inactive, as illustrated by the box
324, but the system 300 will not power down. For example, the OS
power-down prevention engine 310 can capture a final write to a
power management block associate with the first OS to prevent the
first OS from completing the final step or steps of a power-save
process. In this way, the first OS can be deactivated, but the
system will not power down completely, enabling the system to
reboot from memory.
[0045] In the example of FIG. 3, the memory boot engine 312 boots
the second OS from memory, which activates the second OS, as
illustrated by the box 326. The memory boot engine 312, as its name
suggests, performs many of the functions expected of a boot loader,
such as disabling interrupts, disabling paging or remap memory, and
setting up the machine for booting an OS. In addition, the memory
boot engine 312 can save power state (this functionality can also
or instead be attributed to the OS state preservation engine 308),
swap contents in low memory with memory contents that are
associated with the second OS, and set or restore the power state
for the second OS. The memory boot engine 312 can then branch to
the appropriate memory location to run the second OS.
[0046] Since state associated with the first OS may remain in
hardware registers, a boot loader of the memory boot engine 312 can
save the state. In this paper, the OS state preservation engine
308, depending upon context, can be treated as if it includes all
functionality associated with saving state, even if some of the
functionality is logically within the boot loader triggered by the
memory boot engine 312. This is because the OA state preservation
engine 308 conceptually includes all aspects of state preservation,
even if state is saved in different manners, at different times, or
by different processes or processors.
[0047] As used in this paper, booting from memory at least includes
allocating memory for a second OS (using, e.g., the OS memory
allocation engine 304), loading data associated with the second OS
in the allocated memory (using, e.g., the OS memory allocation
engine 304), and branching to a relevant memory location to run the
second OS (using, e.g., the memory boot engine 312). Depending upon
the implementation, a memory boot process can also require swapping
contents in low memory with the allocated memory contents and
setting power state for the second OS.
[0048] In the example of FIG. 3, the memory recycling engine 314
reuses memory that was allocated to the first OS and stores the
contents of memory in an OS memory data store 328. The memory
recycling engine 314 (or at least operation of the memory recycling
engine 314) is optional because a system may have sufficient
resources to keep the first OS in memory.
[0049] In the example of FIG. 3, while the second OS is running,
the context switching layer insertion engine 302 awaits a next
switch instruction. When the next switch instruction is received,
the second OS is now the "previous" OS and the first OS or some
other OS is now the "next" OS.
[0050] FIG. 4 depicts a flowchart 400 of an example of a method for
fast context switching. The flowchart 400 is organized as a
sequence of modules. However, it should be understood that these,
and modules associated with other methods described herein, may be
reordered into different sequences of modules or for parallel
execution.
[0051] In the example of FIG. 4, the flowchart 400 continues at
module 402 with inserting a context switching layer between a first
operating system (OS) and a hardware layer. The context switching
layer can be created, for example, using virtualization features of
a processor. The context switching layer can include, for example,
a minimalist hypervisor.
[0052] In the example of FIG. 4, the flowchart 400 starts at module
404 with allocating memory for a second OS. In a specific
implementation, the allocated memory does not initially include
memory locations that are used by the first OS, such as, for
example, low memory.
[0053] In the example of FIG. 4, the flowchart 400 continues at
module 406 with loading data associated with the second OS into the
allocated memory location. The data can include, for example, an
initial system state for the second OS, or can be loaded from an OS
data store that includes system state associated with the second OS
from when it was previously running on the system.
[0054] In the example of FIG. 4, the flowchart 400 continues at
module 408 with instructing the first OS to enter a power-save
mode. The power-save mode can include, for example, a sleep mode.
Although the first OS is instructed to enter the power-save mode,
unless the OS is configured to know that the intention is to switch
from the first OS to the second OS, the first OS will not
necessarily succeed in completely entering the power save mode.
[0055] In the example of FIG. 4, the flowchart 400 continues at
module 410 with preserving state of the first OS. In a specific
implementation, state can be preserved as the first OS carries out
powering down in accordance with entering a power-save mode. For
example, the first OS can freeze processes and stop device I/O,
while leaving system state in memory (e.g., low memory). The first
OS can also freeze application processors while leaving a single
processor running. Some state, such as power state, might be
preserved later. Saving power state later might be necessary
because power-save functionality normally preserves power state by
leaving the contents of control registers alone, and then using the
values in the control registers when the system comes out of
power-save mode. However, since it is the intention in this method
to switch from the first OS to the second OS, the first OS will
lose the power state when the control registers are overwritten
with the power state of the second OS. For this reason, the power
state can be saved after the power-save mode is (nearly) completed
following interception of a final write to a power management
block.
[0056] In the example of FIG. 4, the flowchart 400 continues at
module 412 with intercepting at the context switching layer a final
write to a power management block associated with the power-save
mode. In an implementation that allows an OS to perform all of the
steps associated with a power-save procedure, then capturing only a
final write, it is possible to insert the context switching layer
between the OS and the hardware without actually modifying the OS.
Thus, the OS can be "agnostic" regarding whether there is context
switching. It may be desirable to capture power state of the first
OS after a final write is intercepted so that resources (e.g.,
control registers) can be overwritten when the second OS takes
over. By capturing the power state of the first OS, the system can
switch back to the first OS at a later time and overwrite the
contents of, e.g., the control registers.
[0057] ACPI specifies an interface of registers. Power state can be
referred to as "ACPI state" for ACPI standards-compliant systems.
For an OS that goes to sleep, the ACPI power state normally is not
changed from the time the OS goes to sleep to the time it awakens
again. Guest OSs would see a fake ACPI state (generally, a complete
software emulation of the ACPI state), which means they would not
overwrite ACPI state of a master OS. However, context switching
would destroy ACPI state. Advantageously, the ACPI state of the
first OS can be identified and saved so that the system can switch
back to the first OS even though the second OS overwrites the ACPI
state.
[0058] In the example of FIG. 4, the flowchart 400 continues at
module 414 with booting the second OS from memory. As has been
stated, state of the first OS can be saved at different times. A
boot loader may disable interrupts and paging or remap memory.
Prior to disabling interrupts and paging or remap memory, it may or
may not be difficult to save power state. So, in at least one
implementation, the power state is saved only after interrupts and
paging or remap memory are disabled. Other tasks that can occur
after the disabling of interrupts and paging or remap memory are
swapping data associated with the first OS with data associated
with the second OS. In at least one implementation, OS data is in
low memory. So in order to run the second OS the contents of low
memory must be moved elsewhere (e.g., by swapping the contents of
low memory with the contents of the allocated memory, storing the
contents of low memory to disk, or moving the contents of low
memory to some other location), and the contents of the allocated
memory associated with the second OS must be moved to low memory
(e.g., by swapping the contents of low memory with the contents of
the allocated memory or overwriting the contents of the low memory
with the contents of the allocated memory). In an implementation
that does not require the use of low memory, it is still likely,
but not a requirement, that at least some memory is reserved for
use by an OS. Also, it may or may not be required that the specific
memory locations used by the first OS be replaced with data
associated with the second OS. Regardless of the
implementation-specific details, some arrangement of memory will be
carried out so as to put the data associated with the second OS in
a place that the system expects when the second OS is running It is
theoretically possible that the arrangement is as simple as
changing the pointer from pointing to a location associated with
the first OS to a location associated with the second OS, though
this is unlikely to be effective using computer systems that are on
the market today, which require that the OS occupy expected memory
locations.
[0059] FIG. 5 depicts an example of a computer system 500. The
system 500 may be a conventional computer system that can be used
as a client computer system, such as a wireless client or a
workstation, or a server computer system. The system 500 includes a
device 502, I/O devices 504, and a display device 506. The device
502 includes a processor 508, a communications interface 510,
memory 512, display controller 514, non-volatile storage 516, I/O
controller 518, clock 522, and radio 524. The device 502 may be
coupled to or include the I/O devices 504 and the display device
506.
[0060] The device 502 interfaces to external systems through the
communications interface 510, which may include a modem or network
interface. It will be appreciated that the communications interface
510 can be considered to be part of the system 500 or a part of the
device 502. The communications interface 510 can be an analog
modem, ISDN modem or terminal adapter, cable modem, token ring IEEE
802.5 interface, Ethernet/IEEE 802.3 interface, wireless 802.11
interface, satellite transmission interface (e.g. "direct PC"),
WiMAX/IEEE 802.16 interface, Bluetooth interface, cellular/mobile
phone interface, third generation (3G) mobile phone interface, code
division multiple access (CDMA) interface, Evolution-Data Optimized
(EVDO) interface, general packet radio service (GPRS) interface,
Enhanced GPRS (EDGE/EGPRS), High-Speed Downlink Packet Access
(HSPDA) interface, or other interfaces for coupling a computer
system to other computer systems.
[0061] The processor 508 may be, for example, a conventional
microprocessor such as an Intel Pentium microprocessor or Motorola
power PC microprocessor. The memory 512 is coupled to the processor
508 by a bus 520. The memory 512 can be Dynamic Random Access
Memory (DRAM) and can also include Static RAM (SRAM). The bus 520
couples the processor 508 to the memory 512, also to the
non-volatile storage 516, to the display controller 514, and to the
I/O controller 518.
[0062] The I/O devices 504 can include a keyboard, disk drives,
printers, a scanner, and other input and output devices, including
a mouse or other pointing device. The display controller 514 may
control in the conventional manner a display on the display device
506, which can be, for example, a cathode ray tube (CRT) or liquid
crystal display (LCD). The display controller 514 and the I/O
controller 518 can be implemented with conventional well known
technology.
[0063] The non-volatile storage 516 is often a magnetic hard disk,
flash memory, an optical disk, or another form of storage for large
amounts of data. Some of this data is often written, by a direct
memory access process, into memory 512 during execution of software
in the device 502. One of skill in the art will immediately
recognize that the terms "machine-readable medium" or
"computer-readable medium" includes any type of storage device that
is accessible by the processor 508.
[0064] Clock 522 can be any kind of oscillating circuit creating an
electrical signal with a precise frequency. In a non-limiting
example, clock 522 could be a crystal oscillator using the
mechanical resonance of vibrating crystal to generate the
electrical signal.
[0065] The radio 524 can include any combination of electronic
components, for example, transistors, resistors and capacitors. The
radio is operable to transmit and/or receive signals.
[0066] The system 500 is one example of many possible computer
systems which have different architectures. For example, personal
computers based on an Intel microprocessor often have multiple
buses, one of which can be an I/O bus for the peripherals and one
that directly connects the processor 508 and the memory 512 (often
referred to as a memory bus). The buses are connected together
through bridge components that perform any necessary translation
due to differing bus protocols.
[0067] Network computers are another type of computer system that
can be used in conjunction with the teachings provided herein.
Network computers do not usually include a hard disk or other mass
storage, and the executable programs are loaded from a network
connection into the memory 512 for execution by the processor 508.
A Web TV system, which is known in the art, is also considered to
be a computer system, but it may lack some of the features shown in
FIG. 5, such as certain input or output devices. A typical computer
system will usually include at least a processor, memory, and a bus
coupling the memory to the processor.
[0068] In addition, the system 500 is controlled by operating
system software which includes a file management system, such as a
disk operating system, which is part of the operating system
software. One example of operating system software with its
associated file management system software is the family of
operating systems known as Windows.RTM. from Microsoft Corporation
of Redmond, Wash., and their associated file management systems.
Another example of operating system software with its associated
file management system software is the Linux operating system and
its associated file management system. The file management system
is typically stored in the non-volatile storage 516 and causes the
processor 508 to execute the various acts required by the operating
system to input and output data and to store data in memory,
including storing files on the non-volatile storage 516.
[0069] Some portions of the detailed description are presented in
terms of algorithms and symbolic representations of operations on
data bits within a computer memory. These algorithmic descriptions
and representations are the means used by those skilled in the data
processing arts to most effectively convey the substance of their
work to others skilled in the art. An algorithm is here, and
generally, understood to be a self-consistent sequence of
operations leading to a desired result. The operations are those
requiring physical manipulations of physical quantities. Usually,
though not necessarily, these quantities take the form of
electrical or magnetic signals capable of being stored,
transferred, combined, compared, and otherwise manipulated. It has
proven convenient at times, principally for reasons of common
usage, to refer to these signals as bits, values, elements,
symbols, characters, terms, numbers, or the like.
[0070] It should be borne in mind, however, that all of these and
similar terms are to be associated with the appropriate physical
quantities and are merely convenient labels applied to these
quantities. Unless specifically stated otherwise as apparent from
the following discussion, it is appreciated that throughout the
description, discussions utilizing terms such as "processing" or
"computing" or "calculating" or "determining" or "displaying" or
the like, refer to the action and processes of a computer system,
or similar electronic computing device, that manipulates and
transforms data represented as physical (electronic) quantities
within the computer system's registers and memories into other data
similarly represented as physical quantities within the computer
system memories or registers or other such information storage,
transmission or display devices.
[0071] The present example also relates to apparatus for performing
the operations herein. This apparatus may be specially constructed
for the required purposes, or it may comprise a general purpose
computer selectively activated or reconfigured by a computer
program stored in the computer. Such a computer program may be
stored in a computer readable storage medium, such as, but is not
limited to, read-only memories (ROMs), random access memories
(RAMs), EPROMs, EEPROMs, flash memory, magnetic or optical cards,
any type of disk including floppy disks, optical disks, CD-ROMs,
and magnetic-optical disks, or any type of media suitable for
storing electronic instructions, and each coupled to a computer
system bus.
[0072] The algorithms and displays presented herein are not
inherently related to any particular computer or other Apparatus.
Various general purpose systems may be used with programs in
accordance with the teachings herein, or it may prove convenient to
construct more specialized Apparatus to perform the required method
steps. The required structure for a variety of these systems will
appear from the description below. In addition, the present example
is not described with reference to any particular programming
language, and various examples may thus be implemented using a
variety of programming languages.
* * * * *