JPS59101015A - Error controlling system - Google Patents

Abstract

PURPOSE:To prevent stored contents of a recording medium having yet a lite time from disappearing and to improve the reliability of a word processor, by successively performing the reading and writing of a document of the word processor by treating the whole document in a bundle. CONSTITUTION:A parity sector (RP) is installed to a group (R1-Rn) of information and the parity sector (Rp) has a result which is obtained by adding bits corresponding to each sector of the group R1-Rn at every bit by modulo 2. When the e-th sector becomes illegible even when a usually performed attempt is made again, the procedure shown in the diagram is executed in the process of the second trial. When the step 63 is excuted, it can be said that the content of Re is restored. In this case, if e=0 in the branched step 61, it means that more than two sectors become illegible in the document and restoration is not possible. In case when e=0 at the branched step 62, it means that the Re can be read when a usually performed attempt is tried again.

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical Field of the Invention] The present invention relates to a method for controlling errors that occur due to the lifespan of parts in an information processing apparatus having a recording medium that should be treated as a consumable item.

[Technical background of the invention]

In recent years, the number of information processing devices such as word processors that use floppy disk devices or cassette tape devices as storage devices has been increasing.

The storage media of these devices, such as floppy disks and cassette tapes, typically have a much shorter lifespan than the disks of general hard disk drives and other devices. Therefore, he told users, ``Floppy disks and cassette tapes are consumable items, and the contents of their memories will eventually become unreadable. Instructions and explanations such as "Let's put it away" are provided in the operating manual.

Next, conventional techniques closely related to the present invention will be explained. 1st
The figure shows an outline of the structure of one document in a word processor document file, and is an example in which a floppy disk is used as a recording medium. As is well known, the physical record format of a floppy disk includes a fixed length record format called a sector, and a cyclic long code (CRC) for detecting errors is added to each sector. FIG. 1 shows that this document file consists of n sectors R1 to Rn, each of which has a cyclic long code (CRC). Although the logical contents of the n physical records include document control information in addition to the document text itself, the details of the logical contents and configuration will be omitted here.

[Problems with background technology]

The above-mentioned instructional explanation that recording media have a limited lifespan and will eventually become unreadable is not generally accepted by users, and it is common for users to continue working for long periods of time without making copies.
An accident occurs in which the medium reaches the end of its lifespan and some of its large storage contents are lost. Conventionally, when a certain sector of a document file becomes unreadable, the unreadability is detected using CRC, but it is not possible to restore the stored contents of the sector that has become unreadable.

[Purpose of the invention]

A first object of the present invention is to provide a means for restoring the stored contents even from a storage medium that has reached the end of its lifespan and is no longer readable, in order to avoid the above-mentioned accident.

A second objective is to provide a means for creating a stable and reliable file system using inexpensive storage media with relatively short lifetimes. Furthermore, a third purpose is to detect that the life of a recording medium on which data is recorded using an error correction code is nearing the end, and to provide the user with an opportunity to take appropriate measures.

[Summary of the invention]

Divide the information into a plurality of equal-length parts, add an error detection code to the kernel equal-length parts, and store the bit-wise modulo 2 accumulation result of the equal-length parts with the information. When detecting an error, a user who knows the error can use the modulo 2 accumulation result to restore the memory contents that have become unreadable.

[Embodiments of the invention]

FIG. 2 shows an outline of the structure of a document file on a floppy disk of a word processor to which an embodiment of the present invention is applied. As in the case of FIG. 1, there are n recording sectors marked with R, - circles in the document file, and each sector is labeled C. The feature in FIG. 2 is that the parity sector (
几p) is provided, and this sector also has cnc attached to the first direction.

The above equation has the result of adding the corresponding bits of each sector from Ra to bit by bit, modulo 2. Usually 1
A sector consists of 256 bytes, and one byte consists of 8 bits. Therefore, $ of the jth byte of the i-th sector
Let the bit value be bIjk and use exclusive OR (E
xo 几) When the operation is represented by ■, the j-th bit bp of the j-th byte of Rp can be expressed by the following formula.

bpjk:bIjk■b2Jk■°゛■b2,6J k
・Track time (1) If you add Rp in this way, R
Even if one sector salt of 1-9 becomes unreadable, the following equation holds, so 几 can be restored.

bsJk ”” IJk■b2jk■・・・■ba-I
Jk■b. Ya1. .

■°°°■b25Jl(■b2116Jk■bpjk・
- River... (2) Here, e is the sector number in which the CRC error occurred.

FIG. 4 is an overall diagram of the configuration to which the method of the present invention is applied.

(1) is the CPU, (2J is the main memory, (3) is the display, (4 is the floppy disk, (5) is the keyboard, (6) is the printer, (and is the bus that connects each of the above.

Next, we will explain how to create 几. In general, a word processor moves the entire specified document in a document file normally recorded on a floppy disk to the document working area, performs processing such as proofreading 2 editing on the working area, and after the processing is completed, transfers the entire document to a document file. Store it in Therefore, first of all, it is necessary to set the 256-byte R1 creation area in the conventionally unused main storage area (Fig. 3), and then perform the procedure shown in Fig. 5 when storing the document in the document file. After completing this procedure, the Rp creation area will be filled with Rp.
will remain. Note that FIG. 5 is a reconstruction of only the portions related to the creation of Box 2 extracted from the processing flow of the word processor.

■ The two steps of raw eggs are added for the purpose of this invention.

Next, explain how to restore an unreadable document.If sector e of a document in a document file becomes unreadable even after the normal retry, perform the steps shown in Figure 6 during the retry process. .

Similarly to Figure 5, Figure 6 shows a procedure in which only the parts related to the restoration stage are reconstructed, and the steps that are not marked are the same as the conventional floppy disk reading routine or error retry routine. It is. The procedure shown in FIG. 6 calculates R as follows. In the homogeneous equation, the subscripts J and k are omitted.

b, −(bρb゛2■...■b6-1■b≦■b6
+1■...■b2,6■bp)■b'6...
...(3) Here b'. is incorrect R6. Since the right-hand side of equation (3) is obtained by repeating the operation {circle around (2)} twice, it is obvious that equation (3) holds true. When step (68) is executed according to the principle of equation (3), the contents of Ro are restored. Same branch (6
If e~0 in 1), it means that two or more sectors in the document are unreadable, and in this case, restoration is impossible. If e = 0 at branch (62), it means that R was read by mistake 9 retries, regardless of the present invention.

In order to confirm that the restoration has been performed without error through the above processing, it is sufficient to perform a CRC check on the restored R1.

However, the CRC for each sector of a floppy disk cannot normally be processed by the user.

During creation, error correction parity records for bird CRCs cannot be created. Therefore, in this embodiment, C of Ro
If an error occurs in the 80th copy, it is impossible to correct the error in the CRC, so if you try to perform the above confirmation, you will get an erroneous result that the confirmation was not possible. In order to prevent this inconvenience, it is necessary to check if the content of the 1st creation area (this is referred to as a syndrome) is incorrect in the final created area according to the procedure shown in FIG. (1)
As is clear from the equation, if there is an error only in the CRC, there is no error in the data, so the syndrome is zero. If a burst error spans both data and syndrome, an error will occur in the confirmation process even if the above is done. It can be presumed that an error occurred in the confirmation process due to the error pattern in the syndrome (if the collection of 1 bits C> is at a certain end of the CRC, due to the above-mentioned reason).

In this way, it is also possible to automatically restore a sector that has become unreadable and continue the process.

However, in that case, the user does not know that one sector has become unreadable. Therefore, it is not necessary to continue using this medium, and as a result, another sector will also have an error that cannot be retried, making it impossible to recover.

In order to avoid such a problem, in this embodiment, the above-mentioned restoration technique is not applied during the normal operation of the word processor, and the processing is temporarily discontinued on a floppy disk that cannot be read by just retrying, assuming that it is capable of being read. At this time, the user is advised by a screen message or instruction manual to run the error recovery utility.

The error recovery utility restores the contents of a floppy disk that has become unreadable and copies it to a new floppy disk. Therefore, the above-mentioned problems can be avoided.

So far, we have been talking about the case where one 几 is added to one document. Since the length of documents varies, a more rational system can be achieved by adding 几 to every fixed number of sectors. Also, set the above constant number to a relatively small number (for example, 15)9, so that it is extremely rare for two sectors to become unreadable, and perform the above restoration process while the word processor is running normally. You can also do this. By doing this, the floppy disk is equipped with an error correction function. In this case, monitor the frequency of incorrect code corrections during normal operation, and if the frequency of incorrect code corrections exceeds a certain level, stop the process or send a warning message. 9, the user is requested to run the error recovery utility.

Although the above description has been made using a word processor as an example, it goes without saying that the present invention is applicable to any information processing apparatus. Further, although the description has been made using a floppy disk as an example, it goes without saying that the present invention can also be applied to a cassette disk.

Furthermore, it is a well-known fact that burst errors, in which a number of erroneous pits are grouped together, usually occur in magnetic recording. Although not shown in Figures 1 and 2, a long gap is provided between sectors to prevent one burst error from spanning two sectors. It is being When one sector is a combination of multiple blocks and a CRC is attached to each block, there is a risk that a single burst error may make multiple blocks unreadable. Divide into odd blocks and apply pari dissectors Rp, , R, to each block. It is better to add

〔Effect of the invention〕

The present invention is as described above, and allows the word processor to sequentially read and write the entire document at once. Therefore, with a small amount of memory (256 bytes) and one increase in input/output operations for the entire document, it is possible to restore data with a high fluctuation rate, protecting the stored contents of a long-life storage medium from loss and increasing reliability. It has the effect of

[Brief explanation of the drawing]

FIG. 1 is a schematic diagram of the structure of one document in a word processor document file, FIG. 2 is a schematic diagram of the structure on a floppy disk of a word processor document file to which an embodiment of the present invention is applied, and FIG. 4 is a diagram showing the parity sector creation area in the main memory of the present invention, FIG. 4 is an overall diagram of the configuration to which the method of the present invention is applied, FIG. 5 is a flowchart showing the procedure for storing in a document file, and FIG. The figure is a flowchart showing a procedure for restoring a document that has become unreadable after retrying. (at~几.): A collection of Q's (100 reports, turtle, horse,...
・R crab equal length part, CRC (error) detection code, modulo 2 accumulation result part in 2-bit units. Agent Patent Attorney Mr. Inoue Figure 1 Figure 2 Figure 3 Figure 4 Figure 5

Claims (3)

[Claims] (1) An information processing device that divides and stores a set of information into a plurality of equal-length parts, which adds an error detection code to each of the equal-length parts and adds a bit-wise modulo of the equal-length parts. 2. The cumulative result is added to the above-mentioned batch of information and stored, and in the first process, the above-mentioned false spot detection code is used to detect false spots, and when a false spot is detected, An error control method characterized in that the user is notified of an error, and the error part is restored using the modulo 2 accumulation result in a second process activated by the user based on the notification. (2) It is characterized by having a physical gap between the equal-length parts that are the targets of modulo 2 accumulation in order to prevent a single burst error or an error that spans two equal-length parts. An erroneous allocation method according to claim 1. (3) The number of equal-length parts to be subjected to modulo 2 accumulation is so small that it is extremely rare for two of them to become unreadable, and the error detection code for each equal-length part is used. Claim 1 or 2, characterized in that when the frequency of occurrence of normal error correction reaches a predetermined limit, the user is notified and the erroneous part is restored using the modulo 2 accumulation result. The error control method described in section.