JPH0373904B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a storage device management method in an electronic computer system.

[Conventional technology]

FIG. 1 shows a conventional file management method for managing space in a magnetic disk device. The space 1 of the magnetic disk device is divided into blocks 2 of a fixed size, and the space of the magnetic disk device is managed using a bit map 3 to determine whether each block is in use or empty. bit map j
The (j=1, 2...n)th bit indicates the usage status of the jth block. If we assume that the block with hatching in Figure 1 is in use, and if we express that the block is in use by 1 and if it is empty by 0, then bitmap 3 will be 101110.
...becomes 0.

On the other hand, the information in the bit map 3 may be destroyed in the following cases. In other words, in order to access a magnetic disk device at high speed, a user program may directly access the magnetic disk device by specifying the sector address of the magnetic disk device without going through the file management program, avoiding the overhead caused by file management. However, if the user accidentally writes different data on the bitmap, or if the bit is written in the register that stores the transfer destination address in the disk controller or in the address line that transmits this address to the memory side. If an error or other abnormality occurs in the calculation of the sector address of the magnetic disk device when the magnetic disk device is accessed due to an error in the CPU, and data is written to the wrong location, the bit map may be destroyed.

Also, due to an error in the system program, when the bitmap is being loaded into main memory,
The information in the bitmap is also destroyed by rewriting the bitmap with machine language instructions.

If bitmap 3 in FIG. 1 is rewritten as bitmap 3 in FIG. 2 due to these reasons, the block 21 in use is rewritten as bitmap 3 in FIG.
Since it is shown as empty above, the next time the file is defined, block 21 will be assigned to this file, and the information in block 21 will be destroyed.

Furthermore, since the empty block 22 is expressed as being in use on the bitmap 3, this block becomes unusable even though it is actually usable.

Conventionally, to detect information corruption in a bit map, a parity check method was used in which a 1-bit check bit was added to detect information corruption, or a parity check method was used to detect information corruption by providing a large number of check bits.
ECC (error correcting code, error check and
methods such as the correction method have been proposed.

[Problem to be solved by the invention]

The data that handles most instructions in computers is in bytes, and the input/output unit in magnetic disk drives is sectors. However, since the above parity check method detects the destruction of 1-bit information, and ECC detects the destruction of 2-bit information, it cannot detect the destruction of information in bytes or sectors. I can't.

It is an object of the present invention to solve the above-mentioned problems, to detect the destruction of information in a bit map in units of bytes or sectors, to prevent file destruction due to the destruction of information in the bit map, and to save space on a magnetic disk device. The object of the present invention is to provide a method for preventing the creation of unusable blocks.

[Means to solve the problem]

In order to achieve the above object, the present invention is characterized in that, in a storage device management system, a redundant bit string is provided for a bit map expressing the usage status of a block, and the redundant bit string is filled with data in advance. When setting a predetermined bit pattern and referring to or updating the bit map, it is determined whether the contents of the redundant bit string match the predetermined bit pattern, and if it does not match the predetermined bit pattern, The problem lies in that it is determined that the information in the bit map is destroyed.

[Effect]

According to the present invention, redundant bits are provided for the bits that display the usage status of a block, a predetermined bit pattern is set in advance in the redundant bit string, and the redundant bit string is used when referring to or updating information on a bitmap. It is determined whether the contents of the bit map match a predetermined bit pattern, and if it does not match the predetermined bit pattern, it is determined that the information in the bit map has been destroyed, so the bit map is mistakenly overwritten with other data. It becomes possible to detect the destruction of a bitmap due to the bitmap being destroyed, to prevent the file from being destroyed due to the destruction of the bitmap, and to prevent the creation of unusable blocks in the space of the magnetic disk device.

〔Example〕

FIG. 4 shows the configuration of a bitmap used in the present invention. Bits 311a to 31 that display the usage status of blocks in the bitmap 31
Redundant bits 312a to 312g are provided for 1g. Such a 2-bit set (for example, 311a
and 312a) are hereinafter referred to as entries.

Embodiments of the present invention will be described below with reference to FIGS. 5, 6, 7, and 8. In FIG. 5, redundant bit strings 312a to 312g all have 1 bits.
Set. The numbers (1, 2, . . . n) assigned sequentially to the entries in the bit map 31 from the beginning are hereinafter referred to as entry numbers.

When the bitmap 31 is read from the magnetic disk device into the main storage device, it is checked whether the redundant bit strings 312a to 312g of the bitmap 31 are all 1s. In order to perform this check at high speed, the following command is used. In other words, as shown in Figure 6, the general register
This instruction checks whether the redundant bit is 1 for the number of entries specified by the general register GRI from the address specified by GRO. If the result is normal, it checks the state of the currently executing program. The zero bit and negative bit of the condition code that indicates whether there is a problem are set to 0, and set to 1 when an abnormality occurs.

The operation of this command will be explained using FIG. 6 as an example. First, Figure 6 shows a bitmap consisting of 4096 entries and a size of 2 sectors (512 bytes/sector), where the second sector of the bitmap is accidentally overwritten with data consisting of all 0s. There is. In other words, from the 2049th entry, which is the first entry of the second sector,
The redundant bits up to the 4096th entry are 0.

FIG. 7 shows the operation logic of this instruction, and FIG. 8 shows the instruction format. First, the instruction format consists of only a 1-byte operation code 601, as shown in FIG.

Next, this operating logic is shown in FIG. 40
0 is data or instructions on main memory, 401 is main memory, 402 is memory address register, 4
03 is a decoder that calculates the start address of the microprogram for this instruction from the operation code of the instruction; 404 is a ROM that stores the microprogram; 405 is a microprogram;
06 is a general register, 407 is a work register, 408 and 409 are selectors, 410 is a literal, that is, a register holding a constant, and 412 is a condition code flag. Steps 501~
Execute the flowchart shown in 510. General registers GR0, GR1 before the instruction is executed.
The start address of the bitmap read into the main memory 401 and the number of entries, 4096, are set in each of the bitmaps.

At step 501, the value of the general register GR1 is updated by the number of entries checked at one time (16 in this embodiment). As a result, the value of GR1 is
It becomes 4080. In step 502, the 4 bytes in the bit map information are x1 x 1 x 1 x 1 in binary representation.
×1×1×1×1×1×1×1×1×1×1×1
×1 (× is 0 or 1 depending on the usage status of the block)
is read from main memory 400. In step 503, the 4-byte literal 410 (binary representation
10101010101010101010101010101010, in hex
AAAAAAAAA (this is called a mask pattern) and the bitmap information read in step 502 are logically ORed. The result is displayed in binary
11111111111111111111111111111111, in hex
It becomes FFFFFFFF. This value is stored in the work register.
Set to WK1. In step 504, it is determined based on the value of the general register GR1 whether the bitmap check has been completed. In this case, the value of general register GR1 is 4080, so the step
Link to 505. In step 505, the value of work register WK1 is displayed in binary.
11111111111111111111111111111111 (this is called a test pattern). In this case, as mentioned earlier, the answer is “Yes” and the step
Link to 506. In step 506, the address of the bitmap is updated by 4 bytes, and in step 501
Link here. Such processing is repeated 128 times. At the 129th time, entry number 2049 starts.
2064 entries are checked. That is, in step 501, the value of the general register GRO becomes 2032, and in step 502, the same value as the above mask pattern is written in the work register in binary representation.
Set to WK1. In step 504, as mentioned earlier, the value of general register GR1 is 2032, so a link is made to step 505. step
Since the determination in step 505 is "No," the process is linked to step 507 and the negative bit of the condition code 412 indicating that the bit map is abnormal is set.
Sets the zero bit and ends the instruction.

In this way, it is possible to detect that the bitmap has been destroyed by repainting one sector (512 bytes) with all 0 information.

In steps 508-509, 16 in steps 501-506
Check the remaining fraction after checking each entry. If there is no fraction, (GR)+16 becomes 0 and no check is performed. step 510
This is the case where all redundant bits are 0, so a normal condition code is set.

As in this example, bitmaps can be destroyed not only by being overwritten with other data on the magnetic disk drive, but also by erroneously rewriting 1, 2, or 4 bytes or bits by instructions in main memory. It happens due to things. Generally, machine language instructions mainly process data in units of 1, 2, or 4 bytes, and there is little possibility that the information in the bitmap itself will be destroyed without rewriting redundant bits. Therefore, this method can detect most of the information corruption in the bitmap.

〔Effect of the invention〕

As explained above, according to the present invention, it is possible to detect destruction of information in a bitmap, except in rare cases where all redundant bits are not destroyed. The data that handles most instructions in computers is in bytes, and the input/output unit in magnetic disk drives is sectors (usually 128 to 512 bytes). If the bitmap information is destroyed due to an error such as overpainting the bitmap with other data, multiple bits of information will be destroyed at the same time. In other words, there is a high possibility that information in redundant bits will be destroyed, so it is necessary to detect destruction of information in most bitmaps and prevent file destruction due to destruction of information in bitmaps.
It has now become possible to provide a method for preventing the creation of unusable blocks in the space of a magnetic disk device.

[Brief explanation of drawings]

Fig. 1 is an explanatory diagram of a conventional bitmap, Fig. 2 is an explanatory diagram of problems when the information in the bitmap is destroyed, Fig. 3 is an explanatory diagram of the configuration of a conventional bitmap, and Fig. 4 is an explanatory diagram of the configuration of a conventional bitmap. FIG. 5 is an explanatory diagram of the bitmap configuration in the embodiment. FIG. 6 is a diagram showing an example of a bitmap for explaining a command to check a bitmap. FIG. 7 is an illustration of a bitmap check instruction. FIG. 8 is a diagram showing an instruction flowchart and the configuration of computer hardware on which machine language instructions operate. FIG. 8 is a diagram showing an instruction format of a bitmap check instruction.

Claims (1)

[Scope of Claims] 1. A management method for a storage device used in a computer system, which corresponds to a bitmap representing the usage status of each block of a storage area of the storage device divided into predetermined blocks. A redundant bit string is provided, predetermined bit pattern information is set in the redundant bit string, and when the bit map is referred to or updated, it is checked whether the contents of the redundant bit string match the predetermined bit pattern information. A storage device management method characterized in that the bit map is determined and, when there is a mismatch, destruction of the bit map is detected. 2 A management method for a storage device used in a computer system, in which a redundant bit string is created in correspondence with a bit map consisting of a bit string representing the usage status of each block of the storage area of the storage device divided into predetermined blocks. and setting predetermined bit pattern information in the redundant bit string in advance such that the bit corresponding to the bit map is 1 and the bit corresponding to the redundant bit string is 0 for the bit map and the redundant bit string. A test pattern in which a bit corresponding to the bit map and a bit corresponding to the redundant bit string are 1 is stored, and when referring to or updating the bit map, the predetermined bit pattern information and the mask are stored. A storage device management method characterized in that it is determined whether a value obtained by calculating a logical sum with a pattern matches the test pattern, and when the check pattern does not match, destruction of the bit map is detected.