JPH0315775B2 - - Google Patents

Description

[Detailed description of the invention]

(a) Technical Field of the Invention The present invention relates to a free space search method for managing the storage status of a direct access storage device connected to a processing device, particularly free space. (b) Prior Art and Problems Direct access storage devices (hereinafter referred to as DASD) connected to processing units store a variety of information, and are also erased if unnecessary.
Therefore, it is necessary to manage the free space on this DASD. For this reason, the storage area on the DASD is divided into blocks of a predetermined size, and a bitmap is used in which each block corresponds to one bit.
The bits for unused blocks are '0' and the bits for used blocks are '1', and the DASD
The usage status is managed. Conventionally, in order to detect the number of consecutive empty spaces, ie, '0's, from this bit map, one byte is sequentially input from the bit map into a register, then shifted to the left, and the '0's and '1's are shifted out. A method has been adopted in which continuous free space is detected by counting the number of spaces. This conventional method of searching for free space requires eight shift instructions and eight compare instructions to examine one byte, and these two instructions alone require 16 steps. Therefore, in order to examine the entire bit map, the number of instruction steps becomes extremely large. Therefore, in the conventional method, there was a problem in that the processing speed of free space management was slow. (c) Purpose of the Present Invention In view of the above-mentioned drawbacks of the conventional art, an object of the present invention is to provide a method for searching for free space, which can search for free space quickly and efficiently. (d) Structure of the invention The present invention is directed to a direct access storage device (DASD).
In a processing device equipped with a bit map that shows the usage status of the storage area, the number of '0's that appear first when looking at one byte from the highest bit, and the bit string of '0's starting from the least significant bit. a first conversion table indicating whether or not it is included in a bit pattern correspondence;
If the highest bit of a byte is '0', the number of consecutive '0's starting from the highest bit is recorded, and if the highest bit is '1', the value that can be used to identify that fact is written as a bit. A second conversion table is provided corresponding to the pattern. However, in the second conversion table above, in a special case where there are eight consecutive '0's starting from the highest bit, that is, when all bits are '0', the converted value is changed to the original '8'.
Change it to '0'. Further, when the highest bit is '1', a value of '8' or higher, for example, is associated with it so that this can be identified. When searching for free space, first, the first conversion table is referred to based on the bit pattern of a predetermined byte of a bit map indicating the usage status of the DASD, and the corresponding conversion value is read out. If the converted value is "0", the same process continues for the next byte. If the converted value is different from "0", processing will be stopped and the address of the byte at that time will be set to '0'.
The number of '0's in that byte and whether or not the '0' bit string includes the LSB are detected from the converted value at that time. If the LSB is included, the second conversion table is subsequently referred to based on the bit pattern of the byte of the address next to the first address, and the corresponding conversion value is read out. If the converted value is "0", the same process continues for the byte at the next address, and if the converted value is different from "0", the process is stopped and the address of the current byte is set to '0'. As the address of the last byte of the bit string of . (e) Examples of the present invention Examples of the present invention will be described in detail below with reference to the drawings. FIG. 1 is a block diagram showing the configuration of a processing device used in an embodiment of the free space search method according to the present invention, and FIG. 2 is a diagram showing the configuration of the first and second conversion tables used in the above embodiment. An example diagram, FIG. 3, is a flowchart showing the empty space search method of this embodiment. In FIG. 1, 1 is a main storage device, and 2 is a conversion table, which includes a first conversion table 2-1 and a second conversion table 2-2. 3 is a central processing unit (CPU), 4 is a main memory control unit, 5 is an instruction control unit, 6 is an arithmetic processing unit, 6-
1 indicates a register, and 7 indicates a direct storage device (DASD). In this embodiment, a bit map indicating the usage status of the DASD is created by referring to the first conversion table 2-1 to determine the address of the first byte where the bit string of '0' starts and the number of '0' in that byte. Then, by referring to the second conversion table, find the address of the last byte where the '0' bit string ends and the number of consecutive '0's starting from the most significant bit (MSB) in that byte. Then, from these values, the total number of consecutive '0's and their positions are known, and an attempt is made to search for free space. The bit pattern of a byte consisting of 8 bits is
There are 256 types. Figures 2a and b are tables showing conversion values for these 256 types of bit patterns, with the upper digits of the byte value expressed in hexadecimal notation on the vertical axis and the lower digits on the horizontal axis. It is attached. First, the first conversion table 2-1 will be explained with reference to FIG. 2a. The first conversion table 2-1 shows the number of '0's that appear first when viewing one byte from the highest bit. For example, the bit string for the hexadecimal byte "CB" is '11001011'. There are two bit strings with one or more consecutive bits of '0'. The first conversion table 2-1 shows the number of '0's appearing first from the MSB in this '0' bit string. Furthermore, the first '0' appearing from this MSB
When the bit string contains the least significant bit (LSB), a value obtained by multiplying the number of '0's by 16 is stored so that it can be easily identified that the bit string contains the LSB. For example, the bit pattern of the byte "80" is '10000000', and the number of '0's is seven. When expressed in hexadecimal, this is "07", but the value multiplied by 16 is "70". This value was created by using '0' in 1 byte.
This is because it is an impossible value to appear as the number of , and it is easy to identify.Moreover, by right-shifting this value four times, the true value can be easily obtained. Therefore, in addition to multiplying by 16,
Methods such as setting the MSB to '1' can be used to indicate that the LSB is included. There are only eight types of bit patterns in which the above '0' bit string includes the LSB, and the figure shows "N x 16".
It is shown as. The basic characteristics of the second conversion table 2-2 are MSB
This is a table showing the number of consecutive '0' bits corresponding to bit patterns. However, all bits are '0', i.e.
The hexadecimal number "00" is originally "8", but it is treated as "0", and if the MSB is '1', that is, "80" or more, the original value is "0". ”, but this is set as “8”. In other words, if MSB is '0', 0 to 7, MSB is '
In the case of 1', 8 is assigned. The reason for this configuration is that in this embodiment, the processing performed using the above two conversion tables continues the same processing for the next byte when the detected conversion value is '0'. , if the value is other than '0', the change value for that byte and the address of that byte are returned and processing is stopped.
This point will be discussed later. Next, the processing of this embodiment using the two conversion tables configured as described above will be explained with reference to FIGS. 2a and 2b and FIG. 3. To simplify the explanation, the command for the conversion process in this embodiment, which is performed with reference to the above two conversion tables, will be referred to as a translation test command. This translation test instruction is executed by specifying the address of the byte to be tested and the conversion table. Therefore, to search for free space on a DASD, first specify the start address of the bit map for managing free space and the first conversion table 2-1, and then
Execute the above translation test command. For example, suppose there is a bitmap as shown in Table 1 below.

[Table] In this bitmap, byte 0 has three consecutive '0's, or empty spaces, on the MSB side, and 14 consecutive '0's, or empty spaces, from byte 1 to byte 3. When a translation test instruction is executed for such a bitmap by specifying the first conversion table, a byte pointer is first set at the first byte of the bitmap [see FIG. 3(1)]. The first byte '11001011' in Table 1 above is CB in hexadecimal, and by referring to the first conversion table shown in Figure 2 a, the converted value is ``2''. See Figure (2)]. In this way, when the converted value is a value other than "0", it indicates that the byte to be inspected contains '0'. In other words, a byte indicating a free area has been detected in the bitmap. If the conversion value is "0", that byte is "0", that is, there is no free area, so the translation test instruction is continued for the next byte. If all 256 bytes are “0” [Figure 3 (3)
], advances the map pointer by 256 [see FIG. 3 (4)], and continues the translation test instruction for the next 256 bytes. If the conversion value is "2" as above, the translation test instruction is stopped and the translation test instruction is checked to see if the '0' bit string in this byte is continuous in the next byte. The judgment is made based on whether or not [see Figure 3 (5)]. In this example, the conversion value is “2” and
Since it is smaller than 16, this "2" is input to the register 6-1 shown in Figure 1, and this value is detected as a value with consecutive '0's, that is, the size of the free space [Figure 3]
(6)], detects the address of this byte, and terminates the process. The value of register 6-1 above, ie, "2", indicates that there are two free areas on the leftmost side of the byte, and that this bit string of '0' does not include the LSB. Indicates that they are not consecutive. If the conversion value is greater than 16, e.g.
In the case of a byte of 11111000 ["F8" in hexadecimal], the converted value is "3×16" from FIG. 2a. In this case, the converted value is divided by 16, and the resulting value "3" is input to the register 6-1 as the number of "0"s. [See Figure 3 (7)]. That is, the value shown as "x16" indicates that the '0' bit string of this byte includes the LSB as described above, that is, the '0' bit string continues in the next byte. Therefore, in this case, it is necessary to examine the next byte, and advance the map pointer by one byte [Figure 3 (8)
], continue the translation test using the second conversion table 2-2. For example, if the aforementioned "F8" is followed by the hexadecimal number "1B" [bit string is '00011011'], Figure 2b
"3" can be found. Since the converted value is not "0", the translation test process stops here [see Figure 3 (10)], and the converted value is "8".
In other words, whether there is a continuous '0' bit string starting from the MSB [see Figure 3 (11)],
Since it is not "8", there is a string of consecutive '0' bits starting from the MSB, and the number of bits is three. Therefore, “3” previously stored in register 6-1
and "3" obtained from the second conversion table to obtain "6" as the number of free bits [see FIG. 3 (13)]. In this case, in the translation test process using the second conversion table, the conversion end byte and conversion start byte are the same, so the difference between the addresses of these two bytes is "0", and as shown in Figure 3 ( The second term in equation 13) is “0”. In the test shown in Figure 3 (11) above, if the converted value is "8", this indicates that there is no consecutive '0' bit string starting from the MSB, so the converted value is replaced with '0'. Refer to FIG. 3 (12)] and perform the processing shown in FIG. 3 (13). If the obtained converted value is "0" in the test shown in Figure 3 (10), this means that all bits of this byte are '0', and there is a possibility that '0' continues in the next byte. Therefore, the translation test process continues for the next byte. When all 256 bytes are converted to "0", the number of '0's for 256 bytes, that is, 256 x 8, is added to the value of register 6-1, and this is converted to register 6-1. [See Figure 3 (14)], advance the map pointer by 256 bytes, and
The following 256 translation tests refer to the second conversion table.
Continue with the bite [see Figure 3 (9)]. As described above, in this embodiment, the first conversion table is used to execute the translation test instruction, find the position of the first byte where the '0' bit string starts and the number of '0's, and If the bit string of '0' contains the LSB, execute the translation test instruction again using the second conversion table from the next byte, detect the byte where the bit string of '0' ends, and use that address and The number of consecutive '0' bits is determined from the MSB of that byte, and from these values, the position of the '0' bit string in the bit map and the number of consecutive '0' bits are detected. Note that this translation test instruction is an instruction generally provided for code conversion. For example, there are two character codes: ASCII code and EBCDIC.
are mainly used, and there is often a need to convert between them. The translation test command is used in such cases, and by specifying a pre-prepared conversion table and the start address of the area to be converted, it is possible to detect the converted value in one byte in one step. can. Furthermore, as mentioned above, if the converted value is "0", this instruction will continue to process the next byte, and if the converted value is "0"
Otherwise, store the address and conversion value of that byte in a register and stop processing. The present invention utilizes the above-mentioned properties of the translation test command, and the first conversion table is designed to indicate the number of '0's included in the bit pattern corresponding to the '0'. When you encounter a byte containing
Stops processing and returns the number of '0's in that byte and the address. Therefore, bytes including '0' can be detected with one instruction. In the second conversion table, when all bits are '0', the conversion value is set to '0', and otherwise it is set to other than '0', so when a translation test is executed with reference to this conversion table, When a byte containing '1' is detected, processing stops. Therefore, the byte where the '0' bit string ends can be detected with one instruction. The reason why we set the conversion value when all bits are '0' as '0' instead of '8' is because in this case there is a possibility that the '0' bit string will continue into the next byte.
This is to continue processing. Furthermore, in this conversion table, if the conversion value is "1" to "7", it indicates that there are consecutive '0's starting from the MSB and that the '0' bit string ends at that byte. . Also, if the MSB is '1', it is a value that does not appear if it contains '0',
The value is set to "8" or more in order to make it possible to stop the process. By doing this, in order to detect a 1-byte value, conventionally it required 16 steps with just the shift instruction and the comparison instruction as mentioned above, but with the present invention, by using two types of conversion tables, A translation test command can be executed in one step, and by using this translation test command, a free area can be detected quickly and easily using the procedure explained with reference to FIG. (f) Effects of the Invention As explained above in detail, the free space search method of the present invention uses two types of conversion tables to search for free space at high speed, and is effective in managing the storage area of DASD. It has many advantages.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the system configuration used in an embodiment of the free space search method of the present invention, FIG. 2 is a diagram showing the structure of the first and second conversion tables of the above embodiment, and FIG. The figure is a flowchart of the empty space search method according to the above embodiment. In the figure, 2 is a conversion table, 2-1 and 2-2 are first and second conversion tables, 3 is a central processing unit, and 6-
1 indicates a register, and 7 indicates a DASD.

Claims (1)

[Scope of Claims] 1. In a processing device equipped with a direct access storage device and a bitmap that indicates the storage area usage status of the direct access storage device as a bit value of '0' or '1' corresponding to the storage area, A method of searching for free space in a storage device, in which the number of consecutive '0's in which the first '0' appearing is one or more when a bit string consisting of 8 bits is viewed from the highest bit side. and a first conversion table that indicates whether or not the bit string of '0' includes the least significant bit, corresponding to the bit pattern; When the number of consecutive '0's from the bit is 8, it is "0", and when there are 1 to 7, it is the number of consecutive '0's from the highest bit, and when the highest bit is '1'. In this case, a second conversion table is provided to indicate this in correspondence with the bit pattern, and the corresponding value is read out from the first conversion table based on the bit pattern of the byte at the specified address of the bit map. Value is “0”
In this case, the process is repeated for the byte at the next address, and if the read value is other than "0", the process is stopped, and the address of the byte is stored as the first address of the byte string containing '0'. , find out from the obtained converted value whether the '0' bit string in that byte includes the least significant bit, and if it does not include the least significant bit, calculate the converted value by calculating the number of consecutive '0's. and if the least significant bit is included, the corresponding value is read from the second conversion table based on the bit pattern of the byte of the address next to the first address, and the read value is "0". In this case, the process is repeated for the byte at the next address, and if the read value is different from "0", the process is stopped and the address of the byte to be inspected at that time is set to the byte string containing '0'. While storing it as the final address, find out the number of '0's in the byte from the converted value for the byte, and based on the number of '0's in the bytes of these starting and final addresses and the difference between the two addresses, A free space search method characterized by: calculating the number of consecutive '0's, and determining the size of the free space and the position of the free space from the obtained result.