JPH09114683A - Control method of computer system - Google Patents

Abstract

(57) Abstract: A computer system capable of running a plurality of OSs on one computer is to reduce the loop time for waiting for the end of communication between the OSs. In a virtual computer system, an OS 210
When the OS 210-1 issues a communication state determination command after the -1 starts communicating with the OS 210-2, it is determined whether the communication is in the end state. And, when the communication is finished, O
The end of communication is reported to S210-1. On the other hand, when the communication is in the execution state, the running of the OS 210-1 is suspended and the OS 2
After 10-2 reports the end of communication to OS 210-1, O
S210-1 is set to the runnable state and the communication execution is reported.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a computer system control method, and more particularly to a computer system control method in which a plurality of operating systems run on one computer.

[0002]

2. Description of the Related Art There is a virtual computer system as a computer system capable of running a plurality of operating systems (OS) on one computer. The virtual computer system is a plurality of logical computers under the control of a virtual computer monitor (VMM) (referred to as a hypervisor program in the above publication) as described in JP-A-64-37636. A virtual computer is generated, and the OS corresponding to each virtual computer is run on one computer in a time-sharing manner. In a virtual computer system, one computer is shared by a plurality of OSs, which causes various factors of performance degradation. For this reason, various technologies have been developed in the past for supporting various performance enhancements.

For example, a dispatch control method for a virtual machine disclosed in Japanese Patent Laid-Open No. 62-163149 discloses one high performance method for a virtual machine having a multiprocessor configuration. According to the above publication, when the first virtual processor of the virtual machine of the multiprocessor configuration detects the memory lock state, the running of the first virtual processor is interrupted to configure the second virtual processor of the same virtual machine. Virtual processors (virtual processors that have the possibility of securing the memory lock) are sequentially run. As a result, the first virtual processor executes a loop based on a lock state determination instruction issued to confirm the lock release state of the memory and acquire the lock (if another virtual processor holds the lock, the loop is executed again until the lock is released. It is possible to divert the time of issuing the lock state determination instruction and the virtual processor that issued the instruction to secure the lock) to the running of another virtual processor, and to effectively use the processor of the computer.

On the other hand, the data processing system described in Japanese Patent Laid-Open No. 6-4490 has one or a plurality of coupling mechanisms (CF) and a plurality of processors (in the above-mentioned publication, a central processing decoder (CP).
Called C)) discloses a data processing system connected via a channel. In this data processing system, the OS on the processor sends the send message (S
The MSG) instruction is used to send a request to the coupling mechanism, and the coupling mechanism processes the request and then reports the execution result to the OS on the processor. Further, the OS on the processor can use the test message (TMSG) instruction to determine whether or not the request is being executed in the coupling mechanism.

[0005]

The above-mentioned Japanese Patent Application Laid-Open No. 62-16 / 1987.
The dispatch control method for virtual machines described in Japanese Patent No. 3149 is very useful for improving the performance of a virtual machine having a multiprocessor configuration. However, the above-mentioned JP-A-6-449
If the data processing system described in Japanese Patent No. 0 is attempted to be realized in one computer by a virtual computer system, there are the following problems.

When the data processing system described in Japanese Patent Laid-Open No. 6-4490 is realized in one computer by a virtual computer system, one or a plurality of logics realized as virtual computers in one computer. There is a general coupling mechanism (virtual coupling mechanism) and a plurality of virtual processors, and each virtual processor communicates with the virtual coupling mechanism.

[0007] Here, after the OS on the virtual processor issues the above-mentioned SMSG instruction, in order to confirm the end of communication, T
It is assumed that a loop by the MSG command (a process of issuing the TMSG command again if the communication is not completed and moving to the next process if the communication is completed) is performed. This communication does not end unless the virtual coupling mechanism runs to process the request in the computer and reports the execution result of the request to the OS on the virtual processor. Therefore, if the virtual coupling mechanism is not running on the processor after the OS on the virtual processor issues the SMSG instruction, this OS is
It is impossible to get out of the loop of the G instruction, which wastes processor time.

Therefore, from the viewpoint of effectively utilizing the processor, there is a problem of how to reduce this useless loop. Further, not only the communication between the OS on the virtual processor and the virtual coupling mechanism, but also in a computer system such as a virtual computer system in which a plurality of OSs run in one computer, communication between the OSs is generally performed. There is a problem of how to reduce useless loops waiting for completion.

An object of the present invention is to solve the problem, and to reduce the loop time for waiting for the end of communication between OSs in a computer system in which a plurality of OSs run in one computer. is there.

[0010]

To achieve the above object, a computer system control method according to the present invention is a computer system control in which a plurality of operating systems (OS) share one computer in a time-division manner. A method of determining whether the communication is in an end state in response to the first OS issuing a communication execution state determination command after the first OS starts communication with the second OS. Whether the communication is completed, the communication completion is reported to the first OS,
When the communication is in the execution state, the first OS is put in a running hold state, and the second OS makes the first OS runable in response to the fact that the second OS reports the end of the communication to the first OS. 1
When the OS restarts, the execution state of the communication is reported.

However, it may be controlled so as to report that the communication is being executed when the traveling of the first OS is resumed.

Further, it may be controlled so as to report the end of communication when the traveling of the first OS is resumed.

Further, when the first OS starts communicating with the second OS, the second OS is controlled so as to be in a runnable state.

Multiple operating systems (O
In a computer system in which S) shares one computer in a time-sharing manner, the first OS of the plurality of OSs starts communication with the second OS, and then the first OS issues a communication execution state determination command. Accordingly, it is determined whether or not the communication is in the end state, the communication end is reported to the first OS when the communication is in the end state, and the first OS is in the running hold state when the communication is in the execution state. In response to the fact that the second OS reports the end of the communication to the first OS, the first OS is set to the runnable state, and the execution state of the communication is reported when the running of the first OS is resumed. By controlling in this way, the loop time for waiting for the end of communication between OSs in a computer system in which a plurality of OSs run in one computer is reduced.

If the execution of the first OS is controlled to report that the communication is being executed when the running of the first OS is resumed, the loop based on the execution status determination command of the first OS issues the execution status determination command twice and then the loop is executed from the loop. You can get out.

Further, when the communication end is controlled to report when the first OS resumes running, in the loop based on the execution state determination command of the first OS, the execution state determination command is issued once and then the loop is exited. be able to.

Further, when the first OS starts communication with the second OS, the second OS is set to the runnable state,
The second OS is surely started, and the second OS is the first OS.
Can be notified to the end of the communication.

[0018]

BEST MODE FOR CARRYING OUT THE INVENTION Two embodiments of the present invention will be described below.
This will be described in detail with reference to the drawings.

1. Configuration of Virtual Computer System FIG. 1 is a configuration diagram of a virtual computer system.

In FIG. 1, 100 is a computer and 110 is a virtual computer monitor. Under control of the virtual computer monitor 110, three virtual computers 200-i (i = 1, 2,
3) is generated, and the corresponding operating system (OS 210-1, O 210-1, O in each virtual computer 200-i is generated.
S210-2, OS210-3) are running in time division.

Further, 300-i (i = 1, 2, 3) is
It is a status table for the virtual machine monitor 110 to manage the status of each OS 210-i. Further, the virtual machine monitor 110 queues the OS state table 300-j (j = 1, 2 in FIG. 1) that can be dispatched to the dispatchable queue 400, and the dispatch hold queue 410 holds dispatch. The state table 300-k (k = 3 in FIG. 1) of the operating OS is queued. The detailed configuration of the state table 300 and how to use the dispatchable queue 400 and the dispatch hold queue 410 will be described in detail later.

The virtual computer monitor 110 manages computer resources including a processor, and uses the status table 300 or the like to schedule and boot an OS. Further, the virtual machine monitor 110 simulates a part of the instruction as is conventionally done.

2. Configuration of Commands and Tables Used for Communication Between OSs FIG. 2 is an explanatory diagram of commands and tables used for communication between OSs.

In FIG. 2, the state table 300-1 corresponding to the OS 210-1 is composed of the following fields.

(A) Request destination field 310: This is a field for holding the identifier of the OS of the communication destination.

(B) Execution flag 320: This field holds information as to whether the OS 210-1 corresponding to the state table 300-1 is communicating with another OS.

(C) Busy wait flag 330: As a result of the OS 210-1 corresponding to the status table 320-1 issuing an execution status determination command during communication execution, the OS 210-
This is a field indicating that the traveling of No. 1 is on hold.

(D) State table pointer 340: FIG.
The two queues shown in (the dispatchable queue 400
And a field for holding the address of the next state table 300 in the dispatch hold queue 410).

(E) Register field 350: O next
This is a field indicating the initial value of the register when S is activated.

Using the status table 300-1, how the virtual computer monitor 110 simulates the three commands of the communication start command, the communication end report command, and the execution state determination command will be described below. The busy wait flag 330 and the state table pointer 340
The usage of will be described in detail later.

(A) Communication Start Command The communication start command is a command for starting communication between the OS that issued the command and the requested OS specified by the command. For example, when the OS 210-1 specifies the content of the request and the request destination and issues a communication start command, control is passed to the virtual computer monitor 110, and the register value of the computer 100 at that time is stored in the state table 300-1. It is saved in the register field 350. The virtual machine monitor 110 stores the request destination specified by the instruction in the request destination field 310 of the status table 300-1 corresponding to the running OS 210-1, and the execution flag 320 indicates that communication is in progress. 'Is stored, and the communication start command is simulated by reporting the request content and request source specified by the command to the request destination OS 210-2.

(B) Communication end report command The communication end report command is a command for the OS that received the request to report the execution result of the request to the OS that issued the communication start command. For example, when the OS 210-2 issues the communication end report command specifying the execution result and the report destination, the control is passed to the virtual computer monitor 110 and the value of the register of the computer 100 at that time is running. No. 2 state table 300-2 is saved in the register field 350. Virtual computer monitor 110
Is the status table 3 corresponding to the report destination specified by the command.
"0" indicating the end of communication in the in-execution flag 320 of 00-1
Is stored, and the communication end report command is simulated by reporting the execution result designated by the command to the OS 210-1 of the report destination.

(C) Execution Status Judgment Command The execution status judgment command is a command for reporting the communication status (communication in progress or communication end) to the OS that issued the command. When the OS 210-1 issues the execution state determination command, control is passed to the virtual computer monitor 110, and the register value of the computer 100 at that time is saved in the register field 350 of the state table 300-1.
The virtual machine monitor 110 reports that the communication is being executed to the OS 210-1 when the value of the execution flag 320 of the status table 300-1 corresponding to the OS 210-1 is “1” indicating that communication is being executed, and the execution flag 320 is being executed. The value of is the end of communication.
When it is 0 ', communication with OS 210-1 is completed (communication is not being executed)
The execution state determination instruction is simulated by reporting

3. First Embodiment of Operation of Virtual Machine System FIG. 3 is a flow chart for explaining the operation of the virtual machine system of the first embodiment.

First, in the virtual computer system, the OS
When the 210-1 issues a communication start command to the OS 210-2 (step 3000), the virtual machine monitor 11
This communication start command is assigned to 0. After this indexing, the virtual machine monitor 110 uses the status table 300-1 to change the OS 210-2 to the OS 210-
OS 210-simulating the communication request from
The state table 300-2 of No. 2 is queued in the dispatchable queue 400 to put the OS 210-2 in a runnable state (step 3010). As a result, the OS 210
-2 receives the communication request from the OS 210-1 (step 3020).

Thereafter, when the OS 210-1 issues a communication execution state determination command (step 3030), this execution state determination command is determined by the virtual machine monitor 110.
After this indexing, the virtual computer monitor 110 uses the status table 300-1 as described above to determine whether or not the communication is completed (step 3040). When the communication is completed, the virtual machine monitor 110 communicates with the OS 210-1. (Step 3050). As a result, the determination in the OS 210-1 as to whether or not the communication has ended is YES (step 306
0), the OS 210-1 executes the following processing (for example, processing of the next communication start command).

On the other hand, when the communication is in the execution state, the virtual computer monitor 110 displays the state table 300-of the OS 210-1.
Show the weight in the busy wait flag 330 of 1 '
1'is set, and the status table 300 of the OS 210-1
-1 is queued in the dispatch hold queue 410 to hold the run (step 3070). Such a state is called a busy wait state.

After that, the OS 210-1 starts running (step 3080), and the OS 210-2 changes to the OS 210-.
Issue a communication end report command to 1 (step 30)
90), the communication end report command is issued to the virtual machine monitor 110. When the communication end report command is issued, the virtual machine monitor 110 simulates the communication end report command using the above-mentioned state table 300 (step 3100) and reports the communication end to the OS 21.
It is determined whether or not the value of the busy wait flag 330 of the 0-1 status table 300-1 is 1 (whether busy wait is in progress) (step 3110). Then, when the value of the busy wait flag 330 is “0” (that is, when the busy wait is not in progress), a runnable OS is selected from the dispatchable queue 400 (step 3130).

On the other hand, when the value of the busy wait flag 330 is "1" (that is, during busy wait), the busy wait flag 330 is set to "0" and the state table 300-1 of the OS 210-1 is dispatched. It is removed from the hold queue 410, queued in the dispatchable queue 400 (step 3120), and a runnable OS is selected from the dispatchable queue 400 (step 3130).

Next, it is judged whether the selected OS is the OS 210-1 that released the busy wait (step 3140).
Then, in the case of the OS 210-1, the communication execution is reported at the time of activation (step 3150), and in the case other than the OS 210-1, the selected OS is activated (step 3160).

As a result, when the OS 210-1 is started, the determination of the communication end in step 3060 is NO, and when the OS 210-1 issues the execution state determination command in step 3030 next, it is determined in step 3060 that the communication end. Then, the OS 210-1 can move to the next processing. In this way, a wasteful loop due to the execution state determination command of the OS 210-1 can be suppressed.

4. Second Embodiment of Operation of Virtual Machine System FIG. 4 is a flow chart for explaining the operation of the virtual machine system of the second embodiment.

The flowcharts of FIGS. 4 and 3 are exactly the same except for the following one point. That is, step 315 of FIG.
In 0, the communication execution is reported to the OS 210-1, but in the corresponding step 3150 ′ of FIG.
Report to 0-1.

Therefore, in the first embodiment, when communication is being executed at the start of the loop by the execution state determination instruction,
The OS 210-1 issues the execution state determination command twice (the communication execution is reported at the first time and the communication end is reported at the second time) and exits from the loop. However, in the second embodiment, the execution state determination command is executed. In the loop by, the OS 210-executes the execution state determination command once (the communication end is reported the first time).
The only difference is that it will only be issued.

As described above, according to the present embodiment, it is possible to reduce the loop time for waiting for the end of the communication between OSs in the virtual computer system.

[0046]

According to the present invention, in a computer system in which a plurality of OSs can run on one computer, the OS
It is possible to reduce the loop time for waiting for the end of communication between the two.

[Brief description of the drawings]

FIG. 1 is a diagram showing a configuration of a virtual computer system.

FIG. 2 is a diagram illustrating instructions and tables used for communication between OSs of the present invention.

FIG. 3 is a diagram showing a flowchart for explaining the operation of the virtual machine system according to the first embodiment of this invention.

FIG. 4 is a diagram showing a flowchart for explaining the operation of the virtual machine system according to the second embodiment of this invention.

[Explanation of symbols]

100 ... Computer, 110 ... Virtual computer monitor, 20
0 ... Virtual machine, 210 ... OS, 300 ... Status table, 400 ... Dispatchable queue,
410 ... Dispatch hold queue.

Claims (4)

[Claims] 1. A method according to claim 1, wherein the plurality of operating systems include:
S) is a method for controlling a computer system in which one computer is shared in a time-sharing manner, wherein the first OS among the plurality of OSs starts communication with the second OS,
In response to the first OS issuing a communication execution state determination command, it is determined whether or not the communication is in an end state, and when the communication is in the end state, the communication end is reported to the first OS. When the communication is in the execution state, the first OS is put in the running hold state, and the second OS reports the end of the communication to the first OS, and then the first OS is set in the runnable state. A method for controlling a computer system, wherein the execution status of the communication is reported when the OS is restarted. 2. The method of controlling a computer system according to claim 1, further comprising reporting that communication is being executed when the first OS is restarted. 3. The computer system control method according to claim 1, wherein the end of communication is reported when the first OS restarts running. 4. When the first OS starts communication with the second OS,
2. The computer system control method according to claim 1, wherein the second OS is set to a runnable state.