JPH05265820A - Input/output processing system - Google Patents

Abstract

PURPOSE:To unnecessitate the storage of a special file in a storage device by directly preparing an (i) node on a memory and performing return after a required parameter such as a main number/sub number or the like is set. CONSTITUTION:It is checked whether a directory dispatched in a step S1 is '/dev' to store the special file or not and when the directory is '/dev', it is checked whether the file name is the special file or not in a step S3. When it is the special file, it is checked whether the call is first one or not in a step S5, and when it is the first call, a table for (i) node is prepared on the empty memory in a step S7. Then, the required value such as the main number/sub number is set, the (i) node is returned, and a counter in the (i) node is turned to '1'. When the (i) node corresponding to the special file is already prepared in the S5, the counter is increased for one while referring to the correspondent (i) node, and the (i) node is returned. When the (i) node is obtained, an open processing is executed.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an input / output processing system in a computer.

[0002]

2. Description of the Related Art FIG. J. O of the computer input / output processing method described by Bach (UNIX is a trademark of AT & T Co.)
It is a figure showing the pen process. Based on this, the processing when the conventional driver is opened will be described.

Conventional processing will be described with reference to the following figures. When the UNIX-OS receives an open system call request, it calls namei in the process indicated by A. In namei, the i-node is get based on the passed path name, and passed to the upper layer. Next, after obtaining the main number and the sub number from the inode in B, the driver corresponding to these numbers is called in B.

FIG. 4 is a diagram for explaining the operation when an open system call request is accepted. In FIG. 4, 1 is a main memory, 2 is an operating system loaded in the main memory 1, and 3 is an operating system. Is also an inode for salary and an inode for file 1 loaded in the main memory 1, 4 is an inode for various I / O devices, 5
Is a salary program and data used for the salary program. Further, 6 is a disk device, 7 is a root directory stored in the disk device 6, 8a, 8b,
Reference numeral 8c is a sub-directory under the root directory 7, 9 is a special file under the sub-directory 8a, and 10 is a disk i-node storing the i-nodes of the files in the sub-directories 8b, 8c.

Next, the operation of the open system call will be described with reference to FIG. For example, if the user executes the payroll program and enters the program name payroll as a command, the operating system 2 must load the payroll program from the subdirectory 8b. In that case, the operating system refers to the disk inode 10, loads the salary inode into the main memory 1, and the salary program is stored in the subdirectory 8b based on the main number and the sub number of the salary inode 3b. Identify that there is. In this way, the salary program is loaded into the main memory 1 and the main memory 1 starts executing the salary program 5a. Next, when the payroll program tries to open the file 1, the operating system 2 issues an open system call for the file 1, searches the disk inode 10 in the same manner as before, and stores the inode for the file 1 in the main memory. 1 and stores it as the file 1 inode 3a. Then, it is known that the file 1 exists in the sub-directory 8c from the stored main number and sub-number of the inode for the file 1, and the open process of open is executed for the file 1. The above is the operation of the open process for the program or file.

Next, the open processing of the I / O device when the I / O device is used will be described. If the operating system wants to open and use tty00, the operating system searches for a special file called tty00 under the subdirectory 8a called / dev, loads this tty00 into the main memory 1, and uses it for tty00. Stored as the inode 4c. Under the subdirectory 8a called / dev, special files corresponding to the drivers of all I / O devices that can use this operating system are stored. In this example, for example, tty.
00 to tty99, or kbd00 to kbd
09, or ws00 to ws99, or fx
Special files from d00 to fxd49 are stored.

This special file, for example, each file has a size of 128 bytes, and the special file identifies the device driver for the I / O device included in the operating system. The main number and the sub number for performing the operation are stored. In this way, when the inode 4c for tty00 is loaded into the main memory 1, when the operating system 2 subsequently accesses the I / O device tty00, it is stored in the inode 4c for tty00. The main number and the sub number for the tty00 device driver are referred to, and the actual device driver is accessed based on the main number and the sub number. For example, if the operating system tries to access an I / O device that does not exist in this subdirectory 8a called / dev, an error will occur. For example, if an open process is attempted to access the disk device fxd50, the open process will result in an error because the special file fxd50 does not exist in the subdirectory 8a in FIG. In this way, the special files corresponding to all the I / O devices usable by the operating system are individually stored in the sub-directory 8a called / dev, and the device driver is identified in the special file. The main number and the sub number for performing the operation are stored, and these are loaded into the main memory 1 by the open processing, so that the operating system 2 accesses the actual device driver by referring to the inode of the main memory. It becomes possible.

Here, the special file is / de
A file corresponding to the device driver under v. The file format is the same as a normal file, but it is determined as a normal file by the value set in the i-mode area of the inode. It The timing for creating and deleting special files is
It is impossible during normal use, and when supporting a new device, in advance, / dev
I will create it below.

[0009]

In the conventional method, each time the resource of the I / O device controlled by the driver is expanded, a special file corresponding to the resource is newly increased. There was a problem that it takes a long time to search when getting an inode.

Further, in the conventional system, the number of special files is large, and it is necessary to store the special files corresponding to each I / O device, so that a storage area is required, and a disk device or the like is required. There is a problem that the capacity of the external storage device is unnecessarily pressed.

The present invention has been made in order to solve the above problems, and when an inode is created in the main memory during the open process, it can be created in a short time and the external device such as a disk device can be used. An object of the present invention is to obtain an input / output processing method that does not overwhelm the storage capacity of the storage device.

[0012]

In order to realize the present invention, for example, in the namei called to get the inode in the open, when the passed path name is a special file, the inode is directly , And returns after setting the required parameters, and has the following elements. (A) A storage step of storing identification information for specifying a resource for input / output processing, (b) an open start step for starting the open processing for the resource, and (c) an open processing started by the open start step. The identification information search step of searching the identification information of the resource targeted by the above from the identification information stored in the storage step, and (d) the resource started by the identification information search step and targeted for the open processing. If the device is a predetermined device, the identification information obtained by any one of the identification information creating step of creating and storing the identification information of the resource, (e) the identification information searching step and the identification information creating step. An open execution process that executes open processing based on the above.

[0013]

According to the present invention, when the path name passed to namei is a special file, the i-node is directly created on the memory by the operation of the identification information creating step, and the main number / sub-number etc. are created. Since the return is performed after setting the necessary parameters of the driver, the driver open can be realized without storing the special file in the storage device in advance.

[0014]

【Example】

Example 1. An embodiment of the present invention will be described below with reference to FIGS. FIG. 1 shows an example of the identification information creating process of the present invention. Nam according to one embodiment of the present invention
It is a flowchart of the function called in ei. 2 shows an example of the identification information searching process of the present invention, which is called in the process A of the open algorithm of FIG. 3 (an example of the open starting process of the present invention).
It is a figure which shows the algorithm of mei. FIG. 1 shows a function called from the namei algorithm shown in FIG.
The flowchart of the function called when the inode is not found when searching for the inode from the path name is shown.

The algorithm shown in FIG. 2 searches the disk inode previously stored in the disk device or the like in the storage step. The location D is specified by inputting the path name. In the process of outputting the i-node of the file, the case where the i-node is not found is shown. In the conventional case, an error that the file has no i-node to be opened is returned. However, in the present invention, the name of FIG.
The function shown in FIG. 1 will be called at the position of D in the algorithm of i. The function shown in FIG. 1 which is called at the location of D returns an error when the inode cannot be found and the one specified in the path name is the device driver of the I / O device. Try not to be. The flow of FIG. 1 will be described below.

For example, with an open system call "
When "/ dev / tty00" is specified, the ope of FIG.
Figure 2 to get inodes in the n algorithm
The namei algorithm shown in is called, but at the point D of this namei algorithm, / dev / t
If there is no file name ty00, the process shown in FIG. 1 is performed.

In FIG. 1, first, in S1, the passed directory stores a special file "/ de.
v ”, and if so, it is checked in S3 whether the file name is a special file. If it is a special file, it is checked in S5 whether it is the first call, and if it is the first time, In S7, a table for the inode is created in the vacant memory, the necessary values such as the main number and the sub number are set there, and the inode is returned. In S5, i corresponding to the special file
If the node has already been created, the corresponding inode is referenced, the counter is incremented by 1, and the inode is returned.

When the inode is obtained, the processes of B and C of the open algorithm of FIG. 3 (an example of the open execution process of the present invention) are executed, and the open process is executed.

In FIG. 1, the counter is set to 1 or is incremented by one because the counter is incremented by one each time it is opened. This is because it is possible to determine how many processes are using the value of the counter. Then, although not shown, when a closing process is executed, this counter is decremented by one each time it is closed, and when the last process is closed, the counter is set to 0.
Set to return to. When it becomes 0, this i
The node will not be used in any process and the operating system will be able to remove this inode from memory. By deleting, the memory can be used for other data areas, other inodes, etc., and the memory can be effectively used.

As described above, according to this embodiment, na
In the mei algorithm, if an i-node cannot be searched from the path, if the path name specifies a special file, the i-node is automatically created in memory. Up to / d
The special file 9 existing under the subdirectory 8a called ev is no longer necessary, and the time required for searching the special file can be cut. In addition, the problem that the capacity of the external storage device is pressed due to the absence of the special file can be solved. In this way, in this embodiment, since the special file unnecessary for the open process for the driver is eliminated and the i-node is automatically generated, the input / output is performed without using the special file. An output processing method can be achieved.

Example 2. In the first embodiment, the case where the counter is set to 1 when creating the i-node has been described, but the value of the counter may be set to 2 when the i-node is newly created. In this way, when the counter is created with 2 from the beginning, the counter is increased or decreased by the opening and closing, and for example, even when all the processes are closed, the counter is set to the minimum value of 1 It does not become 0. Therefore, the operating system considers this inode to be in use,
Do not delete this inode from memory. Therefore, it is confirmed that the inode already exists in the memory when the open process is performed again to access the same device driver.
The processing of 7 and S8 will be executed only once for one inode, and in the subsequent open, S6 will be executed.
You will only run the flow. In this way, the routine of creating a new inode and setting the required value in it is only required once, and the inode is
The process of performing t becomes faster.

Embodiment 3. In the above Examples 1 and 2,
Although the UNIX operating system has been described as an example, the present invention can be applied not only to the UNIX operating system but also to other operating systems.

Example 4. In the above embodiment, the word "i-node" is used, and the device driver is identified using the word "i-node main number / sub-number". This is an example, and in the present invention, identification information for identifying other device drivers may be used instead of the main number / sub number of the i-node.

Example 5. In this embodiment, the case where the device driver is created for each I / O device is shown, but the I / O device is an example of a certain resource, and a routine used when the resource is used. Is a device driver and may use other words.

Example 6. In the above embodiment, the function shown in FIG. 1 is called from the point D in FIG.
It does not matter even if it is called from point E in FIG. E
When called by a point, when the passed path name is a special file, the process of searching the directory indicated by the path name below point E can be omitted, and the inode can be directly obtained. The effect is that the device driver open processing becomes faster.

[0026]

As described above, according to the present invention, even if the identification information for the special resource such as the special file is not actually stored in advance, the I / O for the driver is not stored.
O treatment can be performed.

[Brief description of drawings]

FIG. 1 is a flowchart diagram of a function called in namei according to an embodiment of the present invention.

FIG. 2 is a flow chart of namei.

FIG. 3 is a flowchart when the driver is open.

FIG. 4 is a diagram illustrating an OPEN process.

[Explanation of symbols]

 1 Main Memory 2 Operating System 3-4 Inode 6 Disk Unit 7 Root Directory 8 Subdirectory 10 Disk Inode

Claims (1)

[Claims] 1. An input / output processing method having the following elements: (a) a storage step of storing identification information for specifying a resource for which the input / output processing is performed; and (b) an open start step for starting an open processing for the resource. (C) an identification information retrieval step of retrieving the identification information of the resource that has been activated by the open activation step and has been subjected to the open processing from the identification information stored in the storage step, (d) the identification information An identification information creating step of creating and storing identification information of the resource when the resource that is activated by the search step and is a target of the open process is a predetermined device, (e) the identification information searching step and the identification information An open execution process that executes an open process based on the identification information obtained by any of the creation processes.