JPH05274200A - Purge control method for name cache mechanism in pass name retrieval - Google Patents

Abstract

(57) [Summary] [Object] The present invention relates to a purge control method of a name cache mechanism in a path name search, and an object thereof is to speed up the search of a path name and prevent a decrease in the processing efficiency of the entire system. [Structure] In a multi-layered file system, a system that identifies a file by a path name, and a system that has a name cache mechanism consisting of a name cache table that associates a file name with a file management table in the system The name cache table is purged in a unit of a selectively designated range, for example, a device name or a file system.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for improving the basic operation performance of a system in which a file is specified by specifying a path name in a system having a multi-layered file system.

FIG. 4 is a diagram for explaining a conventional file search method when a specific file is accessed in a data processing system having a conventional multi-layer file system. FIG. A configuration example of a file system is shown, and (b) shows a conventional method for searching the multi-layered file system.

In the multi-layered file system, FIG.
Although it has a hierarchical structure as shown in (a), in the actual file storage device 2, as shown in FIG. 4 (b), a "file management table" (i-node )
20a to 20a are provided in the form of an index list (i list) 20, and when searching for a specific file that constitutes the multi-layered file system of FIG. 4 (a), first,
Refer to the "file management table" (i-node) 20a (this is set in the system-specific position) designated by "/" at the top of the root file system, and "(I-node) 20a reads the directory file 200 pointed to, searches for a file under it, such as" DIR1 ", that is suitable for finding the required file (for example," FILE1 "), Recognize the file number (this number indicates the position of index list 20 above) (R1),
"File management table" (i
-node) Detect 20b.

The "file management table" of the number (i-nod
e) 20b constitutes a directory file 201 for files under the file "DIR1", searches the directory file 201, and locates on the index list 20 corresponding to a predetermined file "FILE1". That is, the number n2 is detected, the "file management table" (i-node) 20c corresponding to the number n2 is detected,
The actual file under the directory file "DIR10" 20c, for example, the above-mentioned "FILE1" and the addresses of other files "FILE2" are recognized, and then the target file "FILE1" is accessed. ..

In this file search method, it is necessary to access the file storage device 2 every time a directory file in one hierarchy of the multi-layered file system is traced, and the path name search efficiency is significantly reduced. was there.

Therefore, the name cache mechanism 1 has been introduced in the improved path name search method. FIG. 5 is a diagram for explaining a path name search method by the name cache mechanism.

The conventional name cache mechanism is shown in FIG.
As shown in (a), a file name 11 and a file management table (called an fct table) 20a in the system corresponding to the file name 11 on the memory in the system 20a
A table (referred to as a name cache table) 10 that associates with each other is stored, and the file name 11 and the file management table 20a to
Pairs are registered in the name cache table 10, and in the subsequent path name search, the file management tables 20a to 20a corresponding to each file name included in the path name are registered in the name cache mechanism 1
It is a mechanism that makes it possible to obtain by searching for 0.

As a result, the path name search can be performed only by scanning the name cache table 10 on the memory, and the number of input / output operations with the real medium during the path name search can be reduced,
The path name search processing can be speeded up.

However, in a system having the multi-layered file system, the file system is conventionally limited to, for example, a disk device, and it is almost impossible to remove the disk device during system operation. It was nice to say that there is no.

Occasionally, when a specific disk device is removed due to a failure or the like, the power supply is cut off, the corresponding disk device is removed, and then the name cache 1 is deleted when the entire data processing system is initialized. Create a new name cache that supports the new system configuration 1
Since it is possible to operate the system by
Purging the name cache 1 every time one file storage was removed did not cause any practical problems,
Originally, since the file storage device was never removed, there was no problem in the method of purging the name cache 1 by removing the file storage device.

However, with the recent introduction of portable floppy disks, the floppy disks are often removed from the system, and the name cache 1 is purged each time. In the conventional method described above, the name cache 1 cannot search the specific file at high speed, so that an effective method of purging the name cache 1 has been required.

[0012]

2. Description of the Related Art FIG. 5 is a diagram for explaining a path search method using a name cache table mechanism.
FIG. 9 is a diagram illustrating a conventional name cache table purging method.

As described above, the conventional name cache mechanism is the file name 11 in the memory in the system and the file management table 2 in the system corresponding to the file name.
The name cache table 10 that correlates 0a to is held, the pair of the file name 11 of the file searched at the end of the first path name search and the file management table 20a to is registered in the name cache 1, and the path name is searched at the subsequent path name search. It is a mechanism that makes it possible to obtain the file management tables 20a to 20a corresponding to the respective file names 11 contained therein by searching the respective name cache tables 10.

As a result, the path name search can be performed only by scanning the name cache table 10 on the memory, and the number of input / output operations with the real medium during the path name search can be reduced.
The path name search processing can be speeded up.

[0015]

The name cache mechanism 1 uses the file name that has been searched at the first path name search.
This is a mechanism for registering a pair of 11 and the file management table 20a to the name cache table 10 and then using each name cache table 10 when searching the path name of the same file thereafter, for speeding up.

At present, as described above, when the file system is no longer used (when the file system is removed) or when the file management table 20a of the system is exhausted (that is, for the name cache 1). Because the file management tables 20a to 20a in the system are occupied), all the name cache tables 10 of the name cache mechanism 1 are purged as shown in FIG. 6a.

Therefore, when a file system, for example, a portable file storage device such as the above-mentioned floppy disk, is frequently installed / removed, all of the name cache table 10 is frequently purged and effective. Will not work.

FIG. 5A shows a state in which all files are registered in the name cache 1. in this case,
For example, the path name search of "/ DIR1 / DIR10 / FILE1" searches the name cache table 10 (
→ →: Fig. 5 (b) shows the operation flow at that time)
The file management table 20c (see FIG. 4) of the target (“FILE1”) can be searched by simply doing
No actual input / output processing with 1 occurs.

However, as shown in FIG. 6A, if the HD2 is removed (shown by the dotted line) and all the name cache tables 10 are deleted, then "/ DIR" is displayed.
When the path name search "1 / DIR10 / FILE1" runs,
As shown in the operation flow of Fig. 6 (b), the deleted HD
I / O processing for the undeleted HD1 to read the file management table even though it is not a file on file 2 {file management table (fct) A, B,
C acquisition processing} occurs, and the processing speed decreases.

That is, as shown in FIGS. 6 (a) and 6 (b), the path name search involves access to the actual input / output device (HD1), resulting in a significant decrease in path name search efficiency. In addition, there is a problem that the processing efficiency of the entire system is reduced.

In view of the above-mentioned conventional drawbacks, the present invention has the following advantages:
An object of the present invention is to provide a purge control method capable of speeding up the search for a path name and suppressing a decrease in the processing efficiency of the entire system.

[0022]

FIGS. 1 and 2 are views showing an embodiment of the present invention, and FIG. 1 (a) shows a procedure for searching a purgeable name cache table.
FIG. 1B shows the operation flow, and FIGS. 2C1 and 2C2 show the state of the name cache after the purge control according to the present invention is performed. The above problems can be solved by the name cache table purge control method configured as follows.

In a multi-layered file system, a system for identifying a file by a path name holds a name cache mechanism 1 consisting of a name cache table 10 that associates a file name 11 with a file management table (fct) 20a in the system. In the system, the name cache table 10 is purged in units of a selectively designated range (for example, device name 12, file system, etc.).

[0024]

In other words, according to the present invention, only the name cache table 10 corresponding to the file included in the file system to be removed (for example, hard disk device: HD2) is targeted for purging, and the effective name cache table 10 is not deleted as much as possible. By doing so, the name cache table 10 once registered is effectively reused, and the processing speed of the path name search is improved.

The name cache table 10 for files to be purged and the name cache table 10 for files not to be purged are separated into the device name 12, as described above.
Alternatively, any information may be used as long as it is information dependent on the file system, such as a device number, and that each file can be recognized as being under the control of that file system.

When the target file system is removed, the name cache table 10 of each file on the file system is compared with the file system identification information (that is, the device name 12, device number, etc.). Then, only the matching name cache table is purged.

FIG. 4A shows the hard disk device H.
It shows a state in which D1 and HD2 are attached to "/ DIR1" and "/ DIR2". In this state, "FILE
1 "," FILE2 "," FILE3 "," FILE "
When a path name search of 4 "occurs, the structure of the name cache 1 of FIG. 4 (a), that is, the name cache table 10 of the paired structure of the file name 11 and the file management table 20a-is generated by the improved conventional method. Created.

Here, when the HD 2 is removed, the structure shown in FIG. 6A is conventionally used, and all the name cache tables 10 on the name cache 1 are purged. Therefore, as described above, in the case of the present invention, FIG.
Name cache 1 according to the operation flow shown in
However, for example, the name cache table 10 corresponding to the file on the removed device (HD2) is searched by the device name 12 and purged (indicated by a circle). As a result, in subsequent path name searches for files under the HD1, no access to the actual input / output device occurs, the search is completed only by scanning the name cache 1, and extra actual input / output processing occurs. As a result, the path name search process can be performed at high speed.

[0029]

Embodiments of the present invention will now be described in detail with reference to the drawings. FIG. 1, FIG. 2, and FIG. 3 described above are views showing an embodiment of the present invention. Here, FIG. 3 shows a method of searching the name cache table.

According to the present invention, in the multi-layered file system, a file is identified by a path name, and a name cache mechanism 1 comprising a file name 11 and a name cache table 10 for associating a file management table 20a ... In the holding system, the name cache table at the time of path name search
A means to perform 10 purges selectively in a specified range, for example, device name 12, file system, etc.
It is a necessary means for carrying out the present invention. The same reference numerals denote the same objects throughout the drawings.

1 to 3 with reference to FIGS.
The purge control method of the name cache mechanism in the path name search of the present invention will be described with reference to FIG. FIG. 1 (a) shows an example of a table structure of a name cache mechanism for realizing the present invention, and FIG. 1 (b) shows an example of a purge control flow.

In FIG. 1A, each name cache table 10 holds not only the file management table 20a corresponding to the file name 11 but also the device name 12 as an identifier of the device in which the file exists. The device name 12
May exist in the so-called file management table 20a.

Here, when purging of the name cache 1 occurs due to removal of the hard disk drive (HD2), etc., the name cache table 10 is searched from the beginning, and the device name 12 of each name cache table 10 and
Compare the device name (HD2) of the detached file system,
Purge only the matching name cache table 10.

That is, as shown in the operation flow of FIG. 1B, in the name cache purging method according to the present invention,
File name as information of the name cache table 10
In addition to 11 and the file management table 20a, the device name 12 is held as the file system identification information. At the timing of purging such as removal, the device name 12 is recognized and the device name 12 is displayed on the removal device (for example, HD2). Only the name cache table 10 corresponding to a certain file can be initialized (purged).

Generally, when the name cache 1 is created, as shown in FIG. 3, each search key, for example, a search key such as a device name is based on a known hash logic (calculation formula for determining the search key). A hash table 3 is created that defines the correspondence with the numbers generated for each.

The created hash table 3 and the name cache table 10 are linked (schedule). Therefore, the name cache table 10
When searching for, the hash table 3 is searched, the number corresponding to the search key is specified, and the name cache table 10 linked to the entry corresponding to the number is searched. The name cache table 10 corresponding to the key can be detected at high speed.

For example, when there are 100 name cache tables 10, by using the hash table 3, it is possible to retrieve the desired name cache by simply searching the hash table 3 composed of, for example, about 10. Table 10 can be detected.

In the above embodiment, the device name 12 is used as the file system identification information, and by comparing this, it is possible to purge only the name cache table 10 of the file on the removed file system.
The present invention is not limited to this, and instead of the device name 12, a device number corresponding to the device name 12 in a one-to-one correspondence is used as a key to search the name cache table 10 that may be purged. It goes without saying that it is okay.

FIG. 2 (a) shows the state of the name cache 1 after the purge control of the name cache table 10 according to the present invention is executed. As is clear from this figure, for example, when removing the hard disk drive (HD2)
The name cache table 10 purged by (shown by the dotted line) is limited to files on the HD2 only,
It can be seen that there is no effect on the name cache table 10 of. That is, "/ DIR1 / DIR1" after that
The path name search of 0 / FILE1 ”(see FIG. 2 (c2), details omitted) is possible without actual input / output processing.

As described above, the present invention is a system in which a file is specified by a path name in a multi-layer file system, and holds a name cache mechanism composed of a name cache table that associates a file name with a file management table in the system. The system is characterized in that the name cache table is purged at the time of searching for a path name in a unit of a selectively designated range, for example, a device name or a file system.

[0041]

As described above in detail, according to the purge control method of the name cache mechanism in the path name search of the present invention, in a multi-layered file system, a file is identified by a path name, In a system that holds a name cache mechanism consisting of a name cache table that associates a file management table in the system, the purging of the name cache table at the time of searching for a path name is performed by selectively specifying a range, for example, device name, Since the file system is used as a unit, for example, it is possible to selectively purge the name cache table within the range limited to the file system (hard disk device etc.) to be removed. In the path name search of the file outside the system To improve the hit rate of over-time cache table, path name search can be performed at high speed. Therefore, in the system operation, in a system in which a path name is frequently searched, there is an effect that the performance of the basic operation of the entire system is improved.

[Brief description of drawings]

FIG. 1 is a diagram showing an embodiment of the present invention (No. 1).

FIG. 2 is a diagram showing an embodiment of the present invention (part 2).

FIG. 3 is a diagram showing an embodiment of the present invention (part 3).

FIG. 4 is a diagram illustrating a conventional file search method.

FIG. 5 is a diagram illustrating a path name search method using a name cache mechanism.

FIG. 6 is a diagram illustrating a conventional name cache table purging method.

[Explanation of symbols]

 1 Name cache mechanism or name cache 10 Name cache table 11 File name 12 Device name 2 i-list 20,20a, 20b, 20c File management table (fct) 200 Root file directory file 201 Directory file 3 Hash table

Claims (1)

[Claims] 1. A multi-layered file system in which a file is identified by a path name, and a name cache table (10) that associates a file name (11) with a file management table (fct) (20a ...) in the system In a system that has the following name cache mechanism (1), the name cache table (10) is purged in units of selectively specified device name, file system, etc. (12) Control method for name cache purging in.