JPH06266535A - Addition system and communication system with addition section - Google Patents

Abstract

PURPOSE:To enable a high-speed processing by providing an addition section successively performing the addition of complement '1' on the read data, compression section performing the addition of complement '1', and output section outputting the compression addition result of the compression section. CONSTITUTION:A 4-byte length addition section 1207 performs the addition of complement '1' by 4-bit length for data to be added in the case the discrimination of the boundary is proved to be 'good' and remaining data to be added in the case of storing save data. After the 4-bit length addition, a 2-byte length compression section 1205 performs the addition of complement '1' by 2-byte length for the 4-byte length data. After the 2-byte length addition, the addition result is held in a 2-byte length data storage area 1209 to perform the discrimination of boundary of the leading data. When the boundary is 'good', the 2-byte length data are stored in an addition result storage area 121 as the addition result and the processing is ended.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a checksum processing addition system used in a network communication system and the like.
The present invention relates to an addition system in which addition is performed by using the complementary addition of 2 to obtain a 2-byte addition result at high speed.

[0002]

2. Description of the Related Art Conventionally, TCP / IP (Transmission
In a network communication system based on the Control Protocol / Internet Protocol), an addition system using 1's complement addition is used for checksum processing performed for data error detection.

First, the network communication system will be described with reference to FIG. FIG. 1 is a diagram showing a network communication system for transmitting and receiving data between two information devices. The transmitter 101 is a workstation,
An information device such as a personal computer, which is a transmission-side user application program 102 for creating transmission data, a transmission-side protocol control unit 103 for controlling a protocol in transmission, and a transmission-side network interface for connecting to a network. Department
Having 105, it transmits data to other information devices. The receiving device 107 is an information device such as a workstation or a personal computer, and a receiving side user application program for processing received data.
It has a receiving side protocol control section 109 for controlling a protocol in receiving, a receiving side network interface section 111 for connecting to a network, and receives data from other information equipment. LAN
A (Local Area Network) 106 transfers data transmitted from the transmission device 101 to the reception device 107, and is typically a FDDI (Fiber Distributed Data Interface).
There is. Addition systems 104 and 110 using 1's complement addition used for checksum processing for error control are performed in a transmission side protocol control unit 103 of a transmission device 101 and a reception side protocol control unit 109 of a reception device 107. ,
The data transmitted from the transmitter 101 is correctly received by the receiver 107.
It is used to check if the received.

A transmission / reception operation will be described with reference to FIG. First, the transmission device 101 passes the transmission data created by the transmission-side user application program 102 to the transmission-side protocol control unit 103, and obtains the addition result as the data to be added of the addition system 104. Then, after adding the obtained addition result to the transmission data, the transmission data is passed to the transmission side network interface unit 105 and transmitted to the reception device 107 via the LAN 106. The reception device 107 includes a reception side network interface unit 11
The received data including the transmission data and the addition result of the addition system 104 on the transmission side received in 1 is passed to the reception side protocol control unit 109. In the receiving side protocol control unit 109,
An addition result is obtained using the received data as the data to be added by the addition system 110, and an error in the received data is detected using the addition result. As a result, when the received data is correct,
Received data is the user application program on the receiving side
It is passed to 108 and is discarded if there is an error.

Next, the flow of the addition processing of the addition systems 104 and 110 used in FIG. 1 will be described with reference to FIG. First, a determination 201 is made as to whether the data to be added is accessible in units of 2 bytes. This is done because the data to be added is arranged on the memory, but the memory has a boundary (hereinafter, referred to as a boundary), and it is impossible to access beyond the boundary. If accessible, SRI (Stanford Research
RFC (Request For Co.) published by Institute
mments) 1071 specifies that the addition byte length of the checksum processing is 2 bytes. Therefore, if the addition byte length is added to the addition data 202, and if access in units of 2 bytes is not possible, 1 byte is added. Long addition 203 is performed.

First, the addition method of the 2-byte length addition 202 will be described with reference to FIG. The data to be added is shown in Fig. 1.
The case where the transmission medium of the LAN 106 of the network communication system shown in FIG. 1 is FDDI and the data is 4470 bytes is taken as an example. 2-byte length addition divides a continuous data string into 2-byte unit data, and
Add using the complement addition of. Added data 301 is 4470
In the case of a byte, the number of additions is 2234, because 2235 pieces of data are added.

Here, an example of addition by a specific numerical value of the 2-byte length addition 202 will be described with reference to FIG. Figure 4
The 16-byte added data 401 in which the start data is stored in an even address is added by 2-byte length addition, and the overflow at the time of addition is not added to the least significant byte every time, but addition processing is performed on the added data. An example is shown in which the addition is performed collectively after the end. The one's complement addition is an operation for re-adding the overflow at the time of addition to the least significant byte, but as shown in FIG. It becomes the same value. First, 2-byte length addition is performed on eight 2-byte length data. At that time, since the overflow 4 of the addition result occurs, it is added to the least significant byte, and the processing is ended. In this case, the addition result 402 is 3BC4 in hexadecimal.

Next, the addition method of the 1-byte length addition 203 will be described with reference to FIG. The data to be added is
As in the case of 4470 bytes of data, as an example. Figure 5
In the case where the boundary is between the odd numbered address and the even numbered address, and the data to be added is stored from the odd numbered address, the memory access cannot be performed in units of 2 bytes. Therefore, as shown in FIG. 3, since it is impossible to perform addition with a 2-byte length, memory access is performed in 1-byte units as shown in FIG. In this case, odd-numbered data is added to the upper byte and even-numbered data is added to the lower byte. At that time, if the addition of the lower byte causes overflow, the addition is performed to the upper byte, and if the addition of the upper byte causes overflow, the addition is performed to the lower byte to obtain the same addition result as the one's complement addition by the length of 2 bytes. I have to. In the 1-byte length addition, a continuous data string is divided into 1-byte units of data, and each of them is added. When the data to be added 501 is 4470 bytes, 4470 pieces of data are added, so the number of additions is 4469.

Here, the addition method of 1's complement addition will be described in detail with reference to FIG. FIG. 6 shows an example in the case of 1-complement addition of 2-digit data A601 and 2-digit data B604. First, the lower digit 602 of data A601 and data B
Add the lower digit 605 of 604 and the lower digit 608 of the addition result 607
Ask for. Next, upper digit 603 of data A601 and data B
The upper digit 606 of 604 is added, and when a digit overflow occurs due to the addition of the lower digit, the overflow is also added to obtain the upper digit 609 of the addition result 607. If a digit overflow occurs due to the addition of the upper digits, the digit overflow 610 is added to the lower digit 608 of the addition result 607, and the one's complement addition is completed. As described above, the one's complement addition is to sequentially add the overflow to the upper digit and add the overflow of the most significant digit to the least significant digit.

[0010]

In the above prior art, when there is a boundary between an odd address and an even address,
If the data to be added is stored from an even address, the addition is performed with a 2-byte length, and if the data to be added is stored from an odd address, the addition is performed with a 1-byte length.

1. If the data length is M bytes, M / 2 additions must be performed, and when M is large, a large number of additions are required.

2. If the data to be added is stored from an odd address, double addition must be performed as in the case of an even address.

Further, the above-mentioned conventional addition system is shown in FIG.
When used in a communication system such as
Since the addition process cannot be performed at high speed, there are the following problems.

3. Since the protocol control unit is provided with the addition system and the protocol buffer for storing the data to be added to perform the addition process, the addition process closes to a large proportion in the process of the protocol control unit, which deteriorates the communication performance.

4. Since the error detection of the received data is performed by the addition processing in the protocol control unit, the error detection cannot be performed at an early stage.

Therefore, the communication system cannot communicate at high speed.

An object of the present invention is to perform high-speed processing in an addition system in order to solve the above problems. Also,
Another object of the present invention is to perform high speed communication in a communication system using an addition system.

[0018]

In order to achieve the above object, there is provided an addition system which reads out data stored in a storage means, performs one's complement addition, and outputs an addition result of a predetermined data length. An addition unit that reads data from the storage unit in a unit larger than the predetermined data length and sequentially performs one's complement addition on the read data; and a predetermined data length for the result of the addition unit. It is achieved by including a compression unit that performs 1's complement addition by dividing the unit into units of, and an output unit that outputs the compression addition result of the compression unit.

Further, a head data judging section for judging whether or not the head section of the data stored in the memory section is the head section of the reading unit of the memory section, and the result of the judgment in the head data judging section, And a data saving unit for holding data from the head portion of the data stored in the memory unit to the break of the head portion of the first read unit of the memory unit when the head unit is not the head unit of the reading unit of the memory unit. The addition unit reads out from the storage unit the data after the break in units of reading by the storage unit when the result of the determination by the head data determination unit is not the beginning portion of the reading unit of the storage unit. , The read data is sequentially subjected to 1's complement addition, and the compression unit determines the result of the determination by the head data determination unit. When the data is not the first part of the reading unit of the storage means, a one's complement addition is performed by dividing the addition result of the data after the break of the adding unit into units of the predetermined data length, and the data saving The data held in the section is further divided into units of the predetermined data length, and one's complement addition is further performed, and the result of compression addition in the compression unit for data after the division is stored in the data saving unit. It further has a save data addition unit for performing one's complement addition of the addition result of the held data in the compression unit.

Further, when the predetermined data length is 2 bytes, the storage means can set the reading unit of the storage means to an even byte of 4 bytes or more. The reading unit of the storage unit may be 4 bytes.

The head data judging section further judges whether the storage address of the head part of the data stored in the storage means is an odd number or an even number, and as a result of the judgment in the head data judging section, the data is stored in the odd number address. In the case where the addition result is added, a data correction unit for performing data exchange between the upper byte and the lower byte of the compression addition result obtained by compressing the addition result of the data after the break in the addition unit by the compression unit.

Further, a communication control unit for controlling transmission / reception of information via a transmission line, a processing unit for processing information, and error detection data for correctly transmitting / receiving transmission information at the time of transmitting / receiving information should be transmitted. An information processing device having a checksum unit for adding an error to information and detecting an error from received information, wherein the checksum unit has storage means for holding the information, and information stored in the storage means. Is provided with addition means for performing 1's complement addition by 4 bytes each, and compression means for performing 1's complement addition by 2 bytes for each addition result, and the addition result of the compression section can be used as the error detection data. it can.

The addition means determines whether or not the head portion of the information held in the storage means is the head portion of the reading unit of the storage means, and if it is not the head portion, it is held in the storage means. The information from the beginning portion of the stored information to the break of the beginning portion of the first reading unit of the storage means is subjected to 1's complement addition with 2 bytes, and the information after the break is added with 4 bytes each from the storage means. Complement addition is performed, and the addition result for the information after the break is further compressed by performing 1's complement addition every 2 bytes, and the compression addition result for the information after the break is obtained.
The addition result of the information up to the break is added and output. The addition means further determines whether the storage address of the head portion of the information held in the storage means is an odd number or an even number. Data can be exchanged between the upper byte and the lower byte, and the exchanged value and the addition result of the information from the leading portion to the break can be added and output.

Further, in a communication system having a transmission line and a plurality of transmission / reception devices for transmitting / receiving information via the transmission line, the transmission / reception device includes a network interface unit for performing transmission / reception processing on the transmission line, and transmission / reception information. To add / detect the checksum data for error detection for correctly transmitting / receiving the transmission information when transmitting / receiving information. And an adder for adding data to be added. The adder divides the addition result into a unit of a predetermined data length for the one-complement addition. Compression means for performing 1's complement addition, and the addition result of the compression section is used for the error detection check. It can be Kusamudeta.

The protocol control section includes the adding section, and, in addition to the processing in accordance with the protocol, instructs the addition section to perform addition data, and the addition section adds processing to the instructed addition data. And outputs checksum data, and the protocol control unit, upon reception, determines whether the received data is correct based on the checksum data in the adding unit. If it is not correct, the received data is discarded.

The protocol control unit transfers the processed information and the management data of the information to the network interface unit at the time of transmission, after the processing according to the protocol is completed, and at the time of reception, the protocol is applied to the received information. The network interface unit includes the addition unit, and the addition unit is provided to the addition unit based on the information transferred from the protocol control unit and the management data of the information at the time of transmission, together with the transmission / reception process. On the other hand, the data to be added is instructed, the checksum data in the adder is added to the information and sent to the transmission path, the information is received from the transmission path at the time of reception, and the adder is based on the received information. To the data to be added, and based on the checksum data in the adder, whether the received data is correct or not. Determines, if correct, the process proceeds to the next processing predetermined, if not correct,
The received data may be discarded.

Further, the network interface unit includes the addition unit, and the protocol control unit instructs the addition unit to be added data at the time of transmission, and adds the checksum data in the addition unit to the information. Then
Upon reception, receiving information from the transmission line, and instructing the adder unit to be added data based on the received information,
At the time of reception, the network interface unit determines whether or not the received data is correct based on the checksum data in the adding unit. If the received data is correct, the process proceeds to a predetermined next process, Is
Received data can be discarded. The protocol control unit includes a common buffer that stores the information, and instructs the addition unit to specify an area for the data to be added and a storage area for the checksum data that are stored in the common buffer. Refers to the area of the data to be added, which is instructed to the protocol control unit, to perform addition,
By storing the checksum data in the designated checksum data storage area, the checksum data is added to the information, and the interface network unit
At the time of transmission, it is possible to instruct the common buffer to output transmission information, and at the time of reception, receive data can be stored in the common buffer.

Further, the network interface unit includes the adding unit, and the adding unit receives transmission information from the protocol control unit at the time of transmission, performs addition processing based on the transmission information, and performs checksum data. And output to the network interface section, and when receiving,
Addition processing is performed based on the information received by the network interface unit to output checksum data, and the received information is transferred to the protocol control unit. The network interface unit receives the checksum data in the addition unit at the time of reception. It is also possible to judge whether or not the received data is correct based on the above. If it is correct, the process proceeds to a predetermined next process, and if not, the received data may be discarded.

The protocol control section transfers the processed information and the management data of the information to the network interface section at the time of transmission, after the processing according to the protocol is completed, and at the time of reception, the received information is processed by the protocol. The network interface unit includes the addition unit, and the addition unit receives transmission information from the protocol control unit at the time of transmission, and performs addition processing based on the transmission information and checks the transmission information. Outputting sum data and transmission information, receiving the information from the transmission path at the time of reception, adding processing based on the information, outputting checksum data, outputting the information to the network interface unit, The interface section, when transmitting, adds checksum data to the transmission information output from the adding section. In addition, the transmission information is sent to the transmission path, and at the time of reception, it is determined whether or not the reception data is correct based on the checksum data in the adding unit. If the reception data is correct, the next predetermined process is performed. Migrated,
If it is incorrect, the received data can be discarded.

The network interface unit includes the adding unit, and the protocol control unit, when transmitting,
Instructing the data to be added to the adder unit, adding the checksum data in the adder unit to the information, and the adder unit receives the information from the transmission path at the time of reception, and performs the addition process based on the information. And output checksum data,
The information is output to the protocol control unit, and when receiving, the network interface unit determines whether the received data is correct based on the checksum data in the adding unit. If it is not correct, the received data may be discarded.

[0031]

When the above means is executed, the head data judging section
It is determined whether or not the address of the head portion of the data stored in the storage means is the head portion (boundary) of the reading unit of the storage means. The data saving unit saves the data from the first data to the accessible break when it is not the first part of the reading unit of the storage unit. Next, the adder reads the data after the break in units of reading from the storage, and sequentially performs one's complement addition on the read data. Further, the compression unit divides the addition result into units of a predetermined data length and performs one's complement addition to generate a compression addition result of the predetermined data length.

Then, if it is the head portion of the reading unit of the storage means, the compression addition result is output and the processing is terminated.

The head data determination unit further determines whether the storage address of the head portion of the data stored in the storage means is odd or even, and the storage address of the head portion of the data stored in the storage means is In the case of an odd address, the data correction unit exchanges the upper byte and the lower byte of the compression addition result. When it is not the first part of the reading unit of the storage means, the save data addition unit adds 1's complement to the save data and the data exchanged data of the data correction unit, outputs the addition result, and the process ends. .

As a result, the data is read from the storage means in a unit larger than the predetermined data length, the one's complement is sequentially added, and the addition result is compressed and added to the predetermined data length. The number of times of addition can be reduced as compared with the conventional number, and the processing time can be shortened.

Further, when the above addition system is used in a communication system, the checksum data can be calculated at high speed, so that the processing in the communication system can be speeded up. In this case, depending on where the addition system is provided in the communication system, various modes can be adopted as described below.

As the communication system, in a communication system having a transmission line and a plurality of transmission / reception devices for transmitting / receiving information via the transmission line, the transmission / reception device includes a network interface unit for performing transmission / reception processing on the transmission line. , A user application unit that processes transmission / reception information, a protocol control unit that processes according to a predetermined protocol at the time of transmission / reception, and at the time of transmission / reception of information,
And an adder unit that adds the data to be added in order to add / detect the checksum data for error detection for correctly transmitting / receiving the transmission information.

When the addition unit is provided in the protocol control unit, the protocol control unit instructs the addition unit to be added data together with the process according to the protocol. On the other hand, the adder performs addition processing on the instructed data to be added, and outputs checksum data. Is correct, and if correct, the process proceeds to the next predetermined process, and if not correct, the received data is discarded.

Further, when the addition unit is provided in the network interface unit, the protocol control unit, at the time of transmission, stores the processed information and the management data of the information after the processing according to the protocol is completed. The information is transferred to the network interface unit, and when received, the received information is processed according to the protocol. At the time of transmission / reception processing, the network interface unit instructs the addition unit based on the information transferred from the protocol control unit and the management data of the information at the time of transmission, and checks in the addition unit. The sum data is added to the information and sent to the transmission path, when the information is received, the information is received from the transmission path, and the added data is instructed to the addition section based on the received information. Based on the checksum data, it is determined whether the received data is correct, and if it is correct, the process moves to the next predetermined process. If it is not correct,
The received data may be discarded.

Furthermore, when the addition unit is provided in the network interface unit, as another aspect,
At the time of transmission, the protocol control unit instructs the addition unit to be added data, adds the checksum data in the addition unit to information, and at the time of reception, receives information from the transmission path and receives the received information. The data to be added is instructed to the adder based on the above. At the time of reception, the network interface unit determines whether or not the received data is correct based on the checksum data in the adding unit. If the received data is correct, the process proceeds to a predetermined next process, Can discard the received data. In this case, the protocol control unit includes a common buffer that stores the information, and instructs the adding unit to specify the area of the added data and the storage area of the checksum data that are stored in the common buffer. The addition unit refers to the area of the data to be added instructed by the protocol control unit, performs addition, and stores the checksum data in the storage area of the instructed checksum data, thereby checking the checksum data. In addition, the interface network unit, when transmitting,
It is possible to instruct the common buffer to output transmission information and store the reception data in the common buffer at the time of reception.

As another aspect of the case where the addition unit is provided in the network interface unit, the addition unit receives the transmission information from the protocol control unit at the time of transmission, and adds based on the transmission information. Processed, add checksum data and output to the network interface unit, and upon reception, perform addition processing based on information received by the network interface unit, output checksum data, and receive to the protocol control unit Transfer information. At the time of reception, the network interface unit determines whether or not the received data is correct based on the checksum data in the adding unit. If the received data is correct, the process proceeds to a predetermined next process, May discard the received data.

Further, as another aspect of the case where the addition unit is provided in the network interface unit, the protocol control unit, at the time of transmission, after the processing according to the protocol is finished, manages the processed information and the information. Data and data are transferred to the network interface unit, and when received, the received information is processed according to the protocol. The addition unit receives transmission information from the protocol control unit during transmission, performs addition processing based on the transmission information, and outputs checksum data and transmission information,
Upon reception, information is received from the transmission path, addition processing is performed based on the information, checksum data is output, and the information is output to the network interface unit. The network interface unit, at the time of transmission, adds checksum data to the transmission information output from the addition unit and sends the transmission information to the transmission path, and at the time of reception, based on the checksum data in the addition unit, It judges whether the received data is correct, and if it is correct, it moves to the next predetermined process, and if it is not correct,
Received data can be discarded.

As another aspect of the case where the addition unit is provided in the network interface unit, the protocol control unit instructs the addition unit at the time of transmission of data to be added and checks the addition unit. Add thumb data to information. Upon receiving, the adder receives information from the transmission path, performs addition processing based on the information, outputs checksum data, and outputs the information to the protocol controller. The network interface unit, when receiving, based on the checksum data in the adding unit, determines whether the received data is correct,
If it is correct, the process moves to the next predetermined process, and if it is not correct, the received data is discarded.

[0043]

EXAMPLE An example of the present invention will be described below. In the present embodiment, the case where the size of data that can be handled by the information processing device that implements the high-speed addition system is 4 bytes is taken as an example.

First, an example in which the high-speed addition system is used for network communication by the TCP / IP protocol in which the transmission medium is FDDI will be described with reference to FIGS. 7 and 8. The TCP / IP protocol has a hierarchical structure including an AP (Application Program), a TCP layer, an IP layer, and a driver layer on both the transmitting side and the receiving side. Is applied. In the process, the high-speed addition system is TC
It is used for the checksum processing performed for data error detection in the P layer and the IP layer.

First, the processing in the information processing device on the transmission side will be described with reference to FIG. FIG. 7 is a diagram showing a hierarchical structure on the transmitting side, a data flow, and a checksum process. On the sending side, first send data with AP701 7
Create 11. In the example shown in FIG. 7, since the maximum packet size on FDDI is 4.5 kilobytes, the transmission data 711 is set to 4430 bytes. The transmission data 711 created by the AP 701 is passed to the TCP layer 702, and in order to improve the accuracy of data error detection, a 20-byte pseudo header 712 added only at the time of checksum in the TCP layer 702,
20-byte TC that stores TCP protocol information
Data for error detection with the P header 713 is added. TC
The checksum unit 705 of the P layer 702 clears the 2-byte TCP checksum value storage area 714 in the TCP header 713 to zero, and then adds the 40-byte header added in the TCP layer 702 and the 4430-byte transmission data 711. 4470 bytes in total are used as the data to be added by the high-speed addition system 706, and the addition result of 2 bytes is obtained. And 1 of the addition result obtained
Of the TCP checksum value storage area 714, and the TCP header 713 and the transmission data 711 are passed to the IP layer 703. The IP layer 703 adds a 20-byte IP header 716 storing the IP protocol information to the passed data. Checksum part of IP layer 703
708 is a 20-byte IP after clearing the 2-byte IP checksum value storage area 715 in the IP header 716 to zero.
A 2-byte addition result is obtained by using the header 716 as the data to be added by the high-speed addition system 709. And the checksum part
703 takes the one's complement of the obtained addition result (processing 71
0), substituting it in the IP checksum value storage area 715, the IP header 716, the TCP header 713, and the transmission data.
Pass 711 to driver layer 704. In the driver layer 704, the 13-byte MAC (Media Access Contr
ol) header and 3-byte LLC (Logical Link Contro)
l) Header and 5-byte SNAP (Sub-Network Access)
Protocol) header with 21 bytes FDDI header
717 is added and the data is transmitted to the receiving side via the FDDI line 718.

Next, the processing in the information processing device on the receiving side will be described with reference to FIG. FIG. 8 is a diagram showing a hierarchical structure on the receiving side, a data flow, and a checksum process. The driver layer 805 has a 21-byte FDDI header 81 transmitted from the transmission side via the FDDI line 801.
4, 20-byte IP header 815, 20-byte TCP header
The FDDI header 814 is removed from the reception data composed of the transmission data 817 of 816 and 4430 bytes, and is passed to the IP layer 804. The checksum portion 810 of the IP layer 804 is the IP header 815.
Is added as data to be added by the high-speed addition system 811 to obtain a 2-byte addition result. Then, the one's complement of the obtained addition result is taken (process 812) and it is judged whether it is 0 (process 813). If it is 0, the received data is correct.
After removing the header 815, the TCP header 816 and the transmission data 817 are passed to the TCP layer 803, and if it is other than 0, the received data has an error and is discarded. In the TCP layer 803, it is added to the passed data only at the time of checksum in the TCP layer 803 in order to improve the accuracy of data error detection.
A 0-byte pseudo header 818 is added.

The checksum unit 809 of the TCP layer 803 uses the pseudo header 818, the TCP header 816, and the transmission data 817 as the data to be added by the high-speed addition system 806 to obtain a 2-byte addition result. Then, the one's complement of the obtained addition result is taken (process 807), and it is determined whether it is 0 (process 8).
08), the received data is correct if 0, so the pseudo header 81
8 and TCP header 816 are removed before sending data 81
7 is passed to AP802, and the reception process is completed. If it is not 0, the received data has an error and is discarded. TCP
In the case of network communication using the / IP protocol, when an error in the received data is detected, the data is discarded and at the same time the sender is requested to send the data again. Is sent.

Next, the characteristics of 1's complement addition used by the high-speed addition system to reduce the number of times of addition processing by 1's complement addition will be described with reference to FIGS. 9 and 10.

First, it will be described with reference to FIG. 9 that the number of digits is not limited in 1's complement addition. FIG. 9 shows an example in which the data C901 and the data D904 having an arbitrary number n digits are subjected to 1's complement addition. First, the least significant digit 902 of data C901 and data D
The lowest digit 905 of 904 is added, and the lowest digit 9 of the addition result 907
Ask for 08. Next, the lower-order second digit 903 of the data C901 and the lower-order second digit 906 of the data D904 are added, and when a digit overflow occurs due to the addition of the lowest digit, the overflow is also added and added. The second lower digit 909 of the result 907 is obtained. Similarly, the addition result is calculated up to the most significant digit, and if an overflow occurs due to the addition of the most significant digit, the overflow 910 is added to the least significant digit 908 of the addition result 907, and 1's complement addition is not performed. finish. As described above, the one's complement addition can be performed by the same procedure even if the number of digits of the data to be added is set to an arbitrary number n. Therefore, by expanding the number of digits of the data to be added to an arbitrary size and performing addition, and further adding the addition result with a 2-byte length, it is possible to obtain the same addition result as the addition with a 2-byte length. is there.

Here, the above addition is expressed as a mathematical expression below.

However, in the following equation, k = 1, ...,
K (K = number of checked sum bytes / n), and “W _{k
(W 1 w 2 ... w n} ) "indicates the content of summing buffer consisting of a value w ₁ w ₂ ... w _n of n bytes," _{U k (u 1 u 2 ...} u n) "
Indicates data consisting of n-byte values u ₁ u ₂ ... u _n ,
“Sum” indicates a checksum value, and “+ ′” indicates 1's complement addition.

[0052]

## EQU1 ## W ₀ = 0

[0053]

[Number 2] _{W k (w 1 w 2 ...} w n) 16 = W k-1 (w 1 w 2 ... w n) 16 + 'U k (u 1 u 2 ... u n) 16

[0054]

## EQU3 ## Sum ₁₆ = W _K (w ₁ w ₂ ) ₁₆ + 'W _K (w ₃ w ₄ ) ₁₆ +' ... + 'W _K (w _n-1 w _n ).
An example of addition of ₁₆ 16-byte data using the case of n = 4 of the above addition will be described with reference to FIG. As shown in FIG. 25, 16-byte data 2501 is first added in 4 bytes. After completion of the addition, as a complement addition of 1, 2 which is an overflow at the time of addition is added to the least significant byte to obtain a 4-byte length addition result 2502. Next, the 4-byte addition result 2502 is added with a 2-byte length. After completion of the addition, as a complement addition of 1, 1 which is an overflow at the time of addition is added to the least significant byte to obtain a 2-byte length addition result 2503. 2-byte length addition result
2503 is 3BC4 in hexadecimal, which is the same as the 2-byte length addition result shown in FIG.

As described above, in the one's complement addition, when there is a digit overflow, the higher digit is added, and the overflow of the highest digit is added to the lowest digit. Therefore, there is no limitation on the number of digits. I understand.

Next, it will be described with reference to FIG. 10 that the addition positions can be exchanged in the 1's complement addition. Figure 10
This is an example of 1's complement addition in which the upper digit and the lower digit of the data to be added are exchanged to perform the addition, and after the addition is completed, the upper digit and the lower digit of the addition result are exchanged. First, the data E1 in which the upper digit 603 and the lower digit 602 of the data A601 as shown in FIG.
Lower digit 1002 of 001, upper digit 606 and lower digit 6 of data B604
The lower digit 1005 of the data F1004 in which 05 is replaced is added, and the lower digit 1008 of the addition result 1007 is obtained. Next, data E10
Add the upper digit 1003 of 01 and the upper digit 1006 of data F1004,
Furthermore, if a digit overflow occurs due to the addition of the lower digits, the overflow is also added and the higher digit 1009 of the addition result 1007 is added.
Ask for. If a digit overflow occurs due to the addition of the high-order digits, the digit overflow 1010 is added to the low-order digit 1008 of the addition result 1007. Then, the higher digit 1009 and the lower digit 1008 of the addition result 1007
By replacing and, the addition result 1011 is obtained, and the 1's complement addition is completed. Lower digit of addition result 1011 after replacement
1012 is the sum of the lower digit 602 of the data A601, the lower digit 605 of the data B604, and the overflow 610 due to the addition of the upper digit, and the upper digit 1013 is the upper digit 603 of the data A601 and the upper digit 606 of the data B604. And the overflow caused by the addition of the lower digits are added, and the same value as the addition result 607 shown in FIG. 6 is obtained. As described above, in the 1's complement addition, the upper digit and the lower digit of the augend data are exchanged to perform the addition, and after the addition is completed, the upper digit and the lower digit of the addition result are exchanged, whereby the augend of the first state is added. It is possible to obtain the same addition result as the addition to the data. Therefore, it is possible to obtain the same addition result as the addition to the original data string by performing the addition from an arbitrary position of the data to be added and correcting the addition position after the end of the addition result.

Here, the above addition is expressed as a mathematical expression below.

However, in the following equation, "W (w
₁ w ₂ ) ”indicates the contents of the addition buffer consisting of the 2-byte value w ₁ w ₂ , and“ U (u ₁ u ₂ ) ”is the 2-byte value u ₁
Data consisting of u ₂ , "Sum" indicates a checksum value, and "+ '" indicates 1's complement addition.

[0059]

[Formula 4] W ₀ (w ₁ w ₂ ) ₁₆ = U ₁ (0u ₁ ) ₁₆

[0060]

## EQU00005 ## W (w ₁ w ₂ ) ₁₆ = W ₀ (w ₁ w ₂ ) ₁₆ + 'U ₂ (u ₂ u ₃ ) ₁₆ +' U ₂ (u ₄ u ₅ ) ₁₆ + '...

[0061]

[Equation 6] Sum ₁₆ = W (w ₂ w ₁ ) ₁₆ 16 bytes of data will be described with reference to FIG. 26 as an example of addition using the above addition. As shown in FIG. 26, first, the 16-byte data 2601 is added by replacing the positions of the first byte data on the right side and the second byte data on the left side. After the addition is completed, as a complement addition of 1, 3 which is an overflow at the time of addition is added to the least significant byte, and an addition result 2602 is obtained. Next, the upper byte and the lower byte of the addition result 2602 are exchanged to obtain the corrected addition result 2603. The addition result 2603 becomes 3BC4 in hexadecimal, as shown in FIG.
It is the same as the 2-byte length addition result shown in.

As described above, in the 1's complement addition, the addition positions are changed and added, and after the addition is completed, the result of further replacement and the result of addition without replacement are the same, so the addition positions can be replaced. Is.

Next, an example of each block constituting the high-speed addition system, processing by each block, and data flow will be described with reference to FIG.

In FIG. 12, the added data storage area
A storage unit 1201 is a place for storing data to be added and is an input unit of the high-speed addition system. The head data determination unit 1202 determines whether or not the head portion of the data to be added is a break in the data length that can be simultaneously read by the storage means. The start data determination unit 1202 determines the storage address of the start data of the data to be added (hereinafter referred to as address determination) and determines whether the top 4 bytes can be accessed at one time (hereinafter referred to as boundary determination). And Boundary determination is to check whether or not there is a boundary in the 4 bytes to be accessed, and if there is no boundary, it is accessible (hereinafter referred to as “boundary suitable”).
If there is, it becomes inaccessible (hereinafter referred to as “unsuitable for boundary”). The data saving unit 1203 saves data from the first data to accessible data when the boundary is inappropriate. The save data storage area 1204 is a place for storing the data saved by the data save unit. The data saving unit 1203 may include a saved data storage area 1204. The 4-byte length addition unit 1207 performs a 1's complement addition with a 4-byte length on the data to be added. Four
The byte length data storage area 1208 is a 4-byte length addition unit 12
It is a place to store 4-byte length data that is the addition result of 07. The 2-byte length compression unit 1205 performs 1's complement addition with 2-byte length on 4-byte length data and save data. 2-byte length data storage area 1209 has 4
This is a place for storing the 2-byte length data that is the result of addition by the 2-byte length compression unit 1205 to the byte length data. The compressed save data storage area 1206 is a place for storing the compressed save data which is the result of addition of the 2-byte length compression unit 1205 to the save data. If the data to be added is stored from an odd address, the data correction unit 1210
The addition result is corrected, and the upper byte and the lower byte of the 2-byte length data are exchanged. The save data addition unit 1211 adds one's complement to the compressed save data and the 2-byte length data when the augend data is not suitable for the boundary. The addition result storage area 1212 is a place for storing 2-byte length data and serves as an output unit of the high-speed addition system. In the above configuration, the 4-byte length addition unit 1207, the 2-byte length compression unit 1205, and the saved data addition unit 1211 may serve as addition units and have all functions. Further, each of the above areas may be configured by one memory by dividing the area on the memory.

Next, the processing flow of the high speed addition system shown in FIG. 12 will be described with reference to FIG.

In FIG. 11, first, the leading data determination unit 1202 determines the boundary of the leading data for the data to be added (process 1101). This is done because the data to be added is located in the memory and there is a boundary in the memory, so access beyond that boundary is not possible. In this embodiment, since 4-byte length data can be accessed, the address at which the top data is stored is detected, and the boundary is determined by determining whether or not this address is the boundary. In addition, depending on the memory, when 8-byte data can be simultaneously accessed, the boundary of 8-byte data may be determined. In this case, the following addition can be performed with 8 bytes.

As a result of the above judgment, if the boundary is unsuitable,
The data saving unit 1203 saves from the head data to accessible data and stores it in the saved data storage area 1204 (process 1102). Then, the 2-byte length compression unit 1205
A 1-complement addition with a 2-byte length is performed on the saved data (process 1103) to generate compressed saved data. For example, when 3 bytes of data are stored in the save data storage area 1204, the lower 2 bytes and the upper 1 byte are added to form a 2-byte length data.

Next, the 4-byte length addition unit 12 is applied to the augend data when the boundary is determined as appropriate in the above determination and the remaining augend data when the saved data is stored.
07 is 4 bytes long and performs 1's complement addition (process 1104),
After the 4-byte length addition is completed, the 2-byte length compression unit 1205 performs a 1's complement addition with a 2-byte length on the 4-byte length data (process 1105). After the addition of the 2-byte length, the addition result is held in the 2-byte length data storage area 1209, and the boundary of the head data is determined (process 1116). This determination is the same as the determination in the above processing 1101, so the determination result may be held at the time of the above determination. If the boundary is suitable,
Addition result storage area with 2-byte length data as addition result
It is stored in 1212 (process 1110) and the process ends.

If the boundary is unsuitable, the head data judgment unit 1202 judges here whether the address of the head data of the data to be added is an odd number or an even number (step 1107). When the first data is stored from the odd address, the 2-byte length compression unit 1205 corrects the 2-byte length data (process 1108). The data correction unit 1210 exchanges the upper byte and the lower byte of the 2-byte length data. When the first data is an odd address, the improper portion is stored in the save data storage area and the remaining data is added. However, as shown in FIGS. Lastly, the upper byte and the lower byte are exchanged, because the addition is performed when they are exchanged. As a result, the data to be added becomes 2
It can be the same as the result of adding byte by byte. If the boundary is unsuitable, the save data addition unit 1211
The compression save data and the 2-byte length data are 1's complement addition (process 1109), the addition result is stored in the addition result storage area 1212 (process 1110), and the process ends.

By processing in this way, it is possible to add at a higher speed.

Next, the hardware configuration of the adder 1207, which is used in the high-speed addition system and performs the addition processing by the one's complement addition, will be described with reference to FIGS. 13 and 24.

First, the case of performing 1's complement addition in two steps will be described with reference to FIG. First, in FIG. 13, each block will be described. The memory 1301 is a storage area for storing the data to be added (in the present embodiment, the data A and the data B are the data to be added) and the addition result. Data bus 1302 is memory 1301
Is a transmission line for exchanging data between the register and the register. The register A 1303 and the register B 1304 are storage areas for temporarily storing the added data transferred from the memory 1301. Switch A 1307 inputs the input of adder 1308 to register A
Switch between 1303 and adder output. Switch B1306
Input the adder 1308 to the register B 1304 or adder 1308
When the addition result of 1 causes overflow, the carry output is switched to the carry output. Adder
1308 outputs the addition result and the overflow at the time of addition by using two pieces of data to be added as input information. The register C1305 is a storage area for storing the output of the adder, and temporarily stores the addition result of 1's complement addition. The control unit 1309
It controls the memory 1301, the register A1303, the register B1304, the register C1305, the switch A1307, the switch B1306, and the adder 1308.

Next, the operation will be described. First, the control unit 1309 uses the data bus 1302 to access the memory 1301.
The data A stored in the register A 1303 and the data B stored in the register B 1304 are transferred to the register B 1304. As step 1,
The control unit 1309 sets the switch A1307 to the register A1303 side,
By switching the switch B1306 to the register B1304 side, the data A and the data B are input to the adder 1308 as the data to be added. When the adding process of the adder 1308 is completed, as a step 2, the control unit 1309 controls the switch A 1307.
To the adder output side and switch B1306 to the carry output side, and the adder 1308 performs the addition again. Then, when the addition processing is completed, since the addition result of the one's complement addition is stored in the register C1305, the control unit 1309
Causes the contents to be transferred to the memory 1301 via the data bus 1302, and the processing ends.

When the data to be added is further present in the memory, the control unit 1309 transfers the data to be added in the memory to the register A 1303 via the data bus 1302 and registers the data stored in the register C 1305. It is transferred to B1304 and operated as in steps 1 and 2 above.

Also, the hardware configurations of the 2-byte compression unit and the save data addition unit are as shown in FIG.
Compression and addition are possible by using a 2-byte register and adder. In case of 2-byte compression unit, control unit
When reading the data to be added from the memory, the reference numeral 1309 reads out and adds the result data obtained by adding 1's complement in 4 bytes in units of 2 bytes. Furthermore, the hardware configuration for data correction is to specify the address to set the data read from the lower byte to the upper byte and the upper byte data to the lower byte when reading the memory. read out.

As described above, the hardware can be used alone and processing can be performed at higher speed.

Next, another configuration of hardware for performing addition processing by one's complement addition will be described. FIG. 24 shows a hardware configuration diagram when the carry output device is used.
A case where one's complement addition is performed in one step by using the carry output device will be described with reference to FIG. First,
Each block will be explained. In FIG. 24, a memory 2401
Is a storage area for storing the data to be added (data A and data B) and the addition result. The data bus 2402 is a transmission path for exchanging data between the memory 2401 and the register. Register A 2403 and register B 2404 are memory
This is a storage area for temporarily storing the data to be added transferred from the 2401. The carry output device 2406 performs addition using the two pieces of data to be added as input information and outputs the overflow at the time of addition. The adder 2407 uses the two added data and the output of the carry output unit 2406 as input information and outputs the addition result. The register C2405 is a storage area for storing the output of the adder, and temporarily stores the addition result of 1's complement addition. The control unit 2408 controls the memory 2401, the register A 2403, the register B 2404, the register C 2405, the carry output device 2406, and the adder 2408.

Next, the operation will be described. First,
The control unit 2408 transfers the data A stored in the memory 2401 to the register A 2403 and the data B stored in the memory 2401 to the register B 2404 via the data bus 2402. Next, the control unit 2408
Data A and data B are input to carry output device 2406 as data to be added. As a result, carry output device 2406
Adds the data A in the register A 2403 and the data B in the register B 2404 and outputs the overflow. Next,
The data A of the register A 2403 and the register B are added to the adder 2407.
The data B of 2404 and the output of the carry output device 2406 are input as the data to be added, and the addition processing is performed. When the addition processing is completed, the addition result of the one's complement addition is stored in the register C2405, so the control unit 2408 transfers the contents to the memory 2401 via the data bus 2402 and the processing is completed.

When there is more data to be added on the memory, the control unit 2408 transfers the data to be added on the memory to the register A 1303 via the data bus 1302, and the data stored in the register C 2405 is registered. Transfer to B2404 and operate as above.

Further, other hardware configurations are also shown in FIG.
Similar to the hardware configuration in 3, the configuration as shown in FIG.

Next, an example in which the addition processing by the 1's complement addition used in the high speed addition system is performed by the C program will be described with reference to FIG. Fig. 14 shows the process of adding the number of bytes to be added in the 4-byte length adder from 4 bytes to 2 bytes, and adding the overflow of the complemented data length = 4470 and 1 to the least significant byte at once after the addition process is completed. The following shows the program to be executed. First, each variable will be explained. The variable data1404 is a 1-byte length array variable, and the data to be added is stored in each element, one byte at a time. The variable * dp1405 is a 2-byte length pointer variable,
It points to the addition data position in the data to be added, and the initial value is the start address of the data to be added. The variable sum1406 is a 4-byte variable and is a storage area in which the addition result of the data to be added is stored, and the initial value is 0. This variable sum1
Since 406 needs to store the addition result of the data to be added, including the overflow, the variable length must be larger than 2, which is the number of bytes added by the 2-byte length addition unit. The variable len1407 is a 4-byte variable, stores the data length to be added, and the initial value is 4470 in FIG.

Next, the flow of processing of this program will be described. First, each variable is declared and initialized (1401). Next, 4470 bytes of the augend data are added with a 2-byte length (1402), and when the addition processing for all the augend data is completed, the overflow from 2 bytes is added 1403 to the least significant byte as a 1's complement addition. . When this program is executed, the addition result of the 2-byte length addition unit is obtained in the variable sum1406. In FIG. 14, the variable su, which is the storage area for the addition result, is
Since the variable length of m1406 was 4 bytes, the number of bytes added in the 2-byte length addition part was 2 bytes so that overflows would not be discarded. In the case where a large value can be obtained, it is possible to increase the addition byte number of the 2-byte length addition unit, that is, 2 as well. For example, if a 4-byte length adding unit is used and the variable length of the variable sum1406 is configured to be larger than 4 bytes, the 4-byte length can be added.

Next, an example of the case where the addition processing by the one's complement addition used in the high speed addition system is performed by the assembler program will be described with reference to FIG. Figure 15
Is a program for performing the process of adding the overflow of the data to be added = 4470, 1's complement to the least significant byte in the process of the 4-byte length adder and the 2-byte length compressor at the same time as the addition process. is there. First, each register will be described. In FIG. 15, the registers used are all 4 bytes in size. regA1504 is an area in which the addition result of the data to be added is stored, and the initial value is 0. The regB 1505 stores the length of the data to be added, and the initial value is 4470 in FIG. regC1506 indicates the addition data position in the data to be added, and the initial value is the start address of the data to be added. regD1507 is an area for storing 4-byte data from the addition data position pointed to by regC1506. regE1508 is an area for storing the lower 2 bytes of the 4-byte length addition result stored in regA1504.

Next, the processing flow of this program will be described. First, each variable is declared and initialized 1501. Next, the data to be added 4470 bytes is added 1502 with a 4-byte length. At this time, the add instruction that adds the overflow of the previous add process at the same time can be used in the assembler language.
The process of adding the overflow to the least significant byte is delayed by one. When the addition of all the data to be added is completed, the overflow of the last addition processing is added to the least significant byte. Next, 2-byte length compression 1503 is performed on the 4-byte length addition result. This is the lower 2 of regA1504
Substitute the byte into regE1508. Next, the value of regA1504 is shifted to the right by 2 bytes to obtain the upper 2 bytes of regA1504. Then, regE1508 is added to regA1504 to perform 2-byte length compression. Finally, since overflow may occur due to 2-byte length compression, the upper 16 bytes and the lower 16 bytes are added as addition of overflow. This ends the processing. When this program is executed, the value obtained by compressing the 4-byte length addition result by 2 bytes in regA1504 is obtained. In FIG. 15, since the register length is 4 bytes, the number of bytes to be added in the 4-byte length addition section is 4 bytes. It is possible to increase the number of bytes to be added, which is 4, to be larger.

Next, in the network communication according to the TCP / IP protocol on FDDI, the addition processing when the high-speed addition system is used for the checksum processing of the TCP layer will be described. Since the number of data bytes that can be handled by the mounted information processing device is 4 bytes, four types of addition processing are performed depending on the boundary in the data to be accessed. In addition, the one's complement addition here is of a form in which overflows at the time of addition are accumulated and finally added together and added to the least significant byte.

First, the 4-byte length addition for the data to be added when the boundary is suitable is shown in FIG. 11, FIG. 16 and FIG.
Will be explained. First, the addition process when the data to be added 1601 is 4470 bytes will be described with reference to FIGS. 11 and 16. First, add 4-byte length to the data to be added 1601
1104 is performed and the addition result is stored in the 4-byte length data storage area. Since the 4-byte length addition 1104 is a 1's complement addition for 1117 4-byte length data and one 2-byte length data, the number of additions is 1117. Next, a 2-byte length compression 1105 is applied to the 4-byte length data of the addition result.
And stores the compression result in the 2-byte length data storage area. Since the 2-byte length compression 1105 is a 1's complement addition for two 2-byte length data, the number of additions is once. Finally, the 2-byte length data is stored 1110 in the addition result storage area, and the process ends.

As described above, the number of additions in the case shown in FIG. 16 is 1118, which is about half the number of additions 2234 in the conventional case shown in FIG.

Next, the adding process for the 16-byte data to be added 1701 with a specific numerical value will be described with reference to FIGS. 11 and 17. First, 4-byte length addition 1104 is performed on four 4-byte length data. At that time, an overflow "2" of the addition result occurs, and the 4-byte length data 1702 is obtained by adding it to the least significant byte. Next, 2 byte length compression for 4 byte length data 1702 11
As 05, the upper 2 bytes and the lower 2 bytes are added. At that time, an overflow "1" occurs in the addition result.
It is added to the least significant byte and the process ends. The addition result 1703 in this case has the same value as the conventional addition result shown in FIG.

Next, the 4-byte length addition is performed on the data to be added whose boundary is not suitable because the storage address of the head data is an odd number and there is a boundary between the first byte and the second byte of the head. This will be described with reference to 18 and FIG. First, the addition process when the data to be added 1801 is 4470 bytes will be described with reference to FIGS. 11 and 18. at first,
The first 1 byte of the data to be added 1801 is stored 1102 in the save data storage area. Then, 2-byte length compression 1103 is performed on the saved data, and the compression result is stored in the compressed saved data storage area. 2-byte length compression 1103 has two 2
Since one's complement addition is performed for byte length data, the number of additions is one. Next, 4-byte length addition 1104 is performed on the second byte of the data to be added 1801 and the addition result is stored in the 4-byte length data storage area. 4-byte length addition 1104 is 1117 4-byte length data and 1 1
Since 1's complement addition is performed for byte length data, the number of additions is 1117. Next, 2-byte length compression 1105 is performed on the 4-byte length data of the addition result, and the compression result is stored in the 2-byte length data storage area. Since the 2-byte length compression 1105 is a 1's complement addition for two pieces of 2-byte length data, the number of additions is one. Next, since the storage address of the head data of the data to be added 1801 is an odd number, data correction 1108 is performed. The data correction 1108 is a byte swap of 2-byte length data, and is performed by exchanging the upper byte and the lower byte. Next, the compressed save data and the 2-byte length data subjected to the data correction 1108 are saved data added 1109. The saved data addition 1109 is added once. Finally, the 2-byte length data is stored 1110 in the addition result storage area, and the process ends.

As described above, the number of additions in this case is 1120.
5 times, about 1/4 of the conventional 4469 additions shown in FIG.
Is.

Next, the adding process for the 16-byte data to be added 1901 using specific numerical values will be described with reference to FIGS. 11 and 19. First, the first byte of the data to be added 1901 is saved 1102. Saved data 1902 compressed by 2 bytes
The compression save data 1903 is obtained by adding the upper 2 bytes and the lower 2 bytes as 1103. Next, 4
Byte length addition 1104 is performed for three 4-byte length data and 3-byte length data in which the vacant lower 1 byte is 0. At that time, since the overflow "2" of the addition result occurs, the 4-byte length data 1904 is obtained by adding it to the least significant byte. Next, 4-byte length data 19
As the 2-byte length compression 1105 for 04, the upper 2 bytes and the lower 2 bytes are added. At that time, since the overflow "1" of the addition result occurs, the 2-byte length data 1905 is obtained by adding it to the least significant byte. Next, since the storage address of the head data of the data to be added 1901 is odd, the 2-byte length data 1905 is set as the data correction 1108.
The correction data 1906 can be obtained by exchanging the upper byte and the lower byte. Finally, the compressed evacuation data
Saved data addition 1109 of 1903 and correction data 1906 is performed.
At that time, since the overflow 1 of the addition result occurs, it is added to the least significant byte, and the process ends. The addition result 1907 in this case has the same value as the conventional addition result shown in FIG.

Next, a 4-byte length addition is performed on the data to be added which is not suitable for the boundary because the storage address of the first data is an even number and there is a boundary between the second and third bytes at the beginning. This will be described with reference to 20 and FIG. First, the addition process when the data to be added 2001 is 4470 bytes will be described with reference to FIG. First, the added data
Store the first 2 bytes of 2001 in the save data storage area 11
02. Then, 2 bytes length compression is applied to the saved data 11
03 is performed and the compression result is stored in the compression save data storage area. Since the 2-byte length compression 1103 is a 1's complement addition for two 2-byte length data, the number of additions is once. Next, 4-byte length addition 1104 is performed from the third byte of the data to be added 2001, and the addition result is stored in the 4-byte length data storage area. 4-byte length addition 1104
Is 1's complement addition for 1117 4-byte length data, the number of additions is 1116. Next, 2-byte length compression 1105 is performed on the 4-byte length data of the addition result, and the compression result is stored in the 2-byte length data storage area. Since the 2-byte length compression 1105 is a 1's complement addition for two pieces of 2-byte length data, the number of additions is one.
Next, the compressed save data and the 2-byte length data are saved data added 1109. The saved data addition 1109 is added once. Finally, the 2-byte length data is stored 1110 in the addition result storage area, and the process ends.

As described above, the number of additions in this case is 1119.
3 times, which is about half of the conventional addition number of 2234 shown in FIG.
Is.

Next, the adding process for the 16-byte data to be added 2101 with a specific numerical value will be described with reference to FIGS. 11 and 21. First, the first 2 of the added data 2101
The byte is saved 1102. The compressed save data 2103 is obtained by adding the high-order 2 bytes and the low-order 2 bytes to the save data 2102 with the 2-byte length compression 1103. Next, add 4-byte length addition 1104 to three 4-byte length data,
This is performed for 2-byte length data in which the vacant lower 2 bytes are set to 0. At that time, an overflow "1" occurs in the addition result, so by adding it to the least significant byte,
The 4-byte length data 2104 is obtained. Next, as the 2-byte length compression 1105 for the 4-byte length data 2104, the upper 2 bytes and the lower 2 bytes are added. At that time, since the overflow "1" of the addition result occurs, the 2-byte length data 2105 is obtained by adding it to the least significant byte.
Finally, the compression save data 2103 and the 2-byte length data 2105 are added. At that time, since a digit overflow "1" occurs in the addition result, it is added to the least significant byte, and the process ends. The addition result 2106 in this case has the same value as the conventional addition result shown in FIG.

Next, the 4-byte length addition is performed on the data to be added which is not suitable for the boundary because the storage address of the head data is an odd number and there is a boundary between the 3rd and 4th bytes at the head. 22 and FIG. 23. First, the addition process when the data to be added 2201 is 4470 bytes will be described with reference to FIGS. 11 and 22. at first,
The top 3 bytes of the data to be added 2201 are stored 1102 in the save data storage area. Then, 2-byte length compression 1103 is performed on the saved data, and the compression result is stored in the compressed saved data storage area. 2-byte length compression 1103 has two 2
Since one's complement addition is performed for byte length data, the number of additions is one. Next, 4-byte length addition 1104 is performed on the data from the 4th byte of the data to be added 2201 and the addition result is stored in the 4-byte length data storage area. Since the 4-byte length addition 1104 is a 1's complement addition for 1116 4-byte length data and one 3-byte length data, the number of additions is 1116. Next, 2-byte length compression 1105 is performed on the 4-byte length data of the addition result, and the compression result is stored in the 2-byte length data storage area. Since the 2-byte length compression 1105 is a 1's complement addition for two 2-byte length data, the number of additions is once. Next, since the storage address of the leading data of the data to be added 2201 is an odd number, data correction 1108 is performed. The data correction 1108 is a byte swap of 2-byte length data, and is performed by exchanging the upper byte and the lower byte. Next, the compressed save data and the 2-byte length data subjected to the data correction 1108 are saved data added 1109. The saved data addition 1109 is added once. Finally 2
The byte length data is stored 1110 in the addition result storage area, and the process ends.

As described above, the number of additions in this case is 1119.
This is about 1/4 of the conventional number of additions of 4469 in FIG.

Next, the addition process for the 16-byte data to be added 2301 with a specific numerical value will be described with reference to FIGS. 11 and 23. First, the first 3 of the added data 2301
The byte is saved 1102. The saved data 2302 is compressed to a 2-byte length 1103, and the upper 2 bytes and the lower 2 bytes are added. At that time, since the overflow 1 of the addition result occurs, the compression save data 2303 is obtained by adding it to the least significant byte. Next, the 4-byte length addition 1104 is performed on the three 4-byte length data and the 1-byte length data in which the vacant lower 3 bytes are set to 0. At this time, since the overflow 1 of the addition result occurs, the 4-byte length data 2304 is obtained by adding it to the least significant byte. Next, 2-byte length compression 1105 for 4-byte length data 2304
By adding the upper 2 bytes and the lower 2 bytes, the 2-byte length data 2305 is obtained. Next, since the storage address of the head data of the augend data 2301 is an odd number, the correction data 2306 is obtained by exchanging the upper byte and the lower byte of the 2-byte length data 2305 as the data correction 1108. Finally, the save data addition 1109 of the compressed save data 2303 and the correction data 2306 is performed. At that time, since the overflow 1 of the addition result occurs, it is added to the least significant byte, and the process ends. Addition result in this case 2307
Becomes the same value as the conventional addition result shown in FIG.

In the present embodiment, the number of added bytes in the 4-byte length adder that performs addition on the data to be added is 4 bytes. Therefore, it is possible to perform the same 2-byte length addition and 8-byte length addition as in the conventional case. By increasing the number of bytes to be added from 4 bytes in this embodiment, the number of times of addition can be further reduced.

In this embodiment, the high-speed addition system is used for the checksum processing performed to detect the error in the received data in the TCP / IP network communication. Although the number is 2 bytes, the addition result can be set to an arbitrary size by expanding the number of addition bytes from 2 bytes to an arbitrary value.

Although the present embodiment has been described as an example in which it is used in the network communication system, it can be similarly used in a system other than the network communication system as long as it is a process of adding continuous data.

According to this embodiment, the number of times of addition can be reduced by performing the addition using the high speed addition system. If the data to be added is M bytes and the number of bytes to be added to the data to be added is 4 (1) If the data to be added is stored from an even address, the length of the data to be added M is 4 bytes. When it is sufficiently larger than the number 4, the number of additions is reduced to 1/2.

(2) If the data to be added is stored from an odd address, and the length of the data to be added M is sufficiently larger than the number of bytes to be added 4 of the 4-byte length adder, the number of additions is 1
It is reduced to / 4.

When the added data length M is sufficiently larger than the number of added bytes 4 of the 4-byte length adder, the number of additions is constant regardless of the storage state of the added data, that is, M / 4 times. There is no difference in the number of additions depending on the storage state. In addition, 4 which is the addition byte length of the 4-byte length addition is
It can be expanded to an even number of bytes according to the size of data that can be handled by the mounted device. For example, when the number of bytes is set to 8 bytes, the number of additions is further reduced.

Next, an embodiment in which the above-mentioned high-speed addition system is applied to a network communication system will be described.
In a network communication system, an example of a high-speed communication system when the above high-speed addition system is provided is shown in FIG. 27 and FIG.
It will be described with reference to FIGS. 28, 29, 30, 31, and 32.
FIG. 27, FIG. 28, FIG. 29, FIG. 30, FIG. 31, and FIG. 32 are diagrams showing a communication system for performing data transfer between two information devices via a LAN whose transmission medium is FDDI. The protocol is TCP / IP. In the embodiments described below, in a communication system having a transmission line and a plurality of transmission / reception devices that transmit and receive information via the transmission line, the transmission / reception device includes a network interface unit that performs transmission / reception processing with respect to the transmission line, A user application unit that processes transmission / reception information, a protocol control unit that processes according to a predetermined protocol at the time of transmission / reception,
When transmitting / receiving information, an adder unit that adds the data to be added in order to add / detect the checksum data for error detection for correctly transmitting / receiving the transmission information is included. The addition unit includes an addition unit that sequentially performs one's complement addition on the information, and a compression unit that divides the addition result into units of a predetermined data length and performs one's complement addition. The addition result is used as the checksum data for error detection.

First, description will be made with reference to FIG. FIG. 27
Shows a configuration diagram in the case where a protocol control unit including a TCP layer, an IP layer, and a driver layer is provided with the above-described high-speed addition system 2709/2712 and a protocol buffer 2708/2711 for storing data to be added. A case where the header addition process and the addition process in the control unit are performed independently will be described. In the present embodiment, the protocol control unit includes the addition unit, and instructs the addition unit to perform augend data along with the processing in accordance with the protocol, and the addition unit determines the addition data to be instructed. Addition processing is performed, and checksum data is output. Upon reception, the protocol control unit determines whether or not the received data is correct based on the checksum data in the addition unit. If it is not correct, the received data can be discarded.

In the protocol control section shown in FIGS. 7 and 8, the header processing and the addition processing are performed together. In the present embodiment, by providing a high-speed addition system, it is possible to perform processing separately for each function.

First, processing on the transmission side will be described. Since the maximum data length permitted on FDDI is 4.5 kilobytes,
Generates 430 bytes of transmit data 2714. Then, the transmission data 2714 is copied to the protocol buffer 2708 provided in the protocol control unit 2702, and the processing of the user application unit 2701 ends. Protocol control unit 2702
Then, first, the TCP layer includes a 20-byte pseudo header 2715 added only in the TCP layer to improve the accuracy of error detection of transfer data, and a 20-byte TCP header 2716 in which TCP protocol information is stored. , Is added before the copied transmission data 2714. And high-speed addition system 27
09 is pseudo header 2715, TCP header 2716 and data
A checksum value is obtained by using 2714 as data to be added, and the checksum value is assigned to the TCP header 2716. Then, the pseudo header 2715 is removed, and the TCP header 2716 and the data 2714 are
And are copied to the IP layer, and the processing of the TCP layer ends. Next, the IP layer stores IP protocol information.
The 20-byte IP header 2717 is added to the T of the copied data.
It is added before the CP header 2716. Then, the high-speed addition system 2709 obtains a checksum value using the IP header 2717 as data to be added, and substitutes the checksum value into the IP header 2717. Then, the IP header 2717, the TCP header 2716, and the data 2714 are copied to the driver layer, and the processing in the IP layer ends. Next, the driver layer adds a 22-byte FDDI header 2718 storing physical address information to the front of the IP header 2717 of the copied data to create a transmission frame 2719. Then, the transmission frame 2719
Copying to the network buffer 2710 provided in the network interface unit 2703, the processing of the driver layer ends. Finally, the network interface unit 27
03 sends the transmission frame 2719 to the LAN 2707, and the processing on the transmission side ends.

Next, the processing on the receiving side will be described. First, the network interface unit 2706 uses the LAN 2707
FDDI header 2722 sent from the sending side via
The received frame 2721 including the IP header 2723, the TCP header 272, and the data 2725 is transferred to the network buffer 2713.
To receive. Next, the network interface unit 27
06 is a protocol control unit 2705 for all received frames 2721.
To the protocol buffer 2711 provided in the. In the protocol control unit 2705, first, the driver layer
The copied data 2721 is processed, and when the processing is completed, the FDDI header 2722 is removed, and the IP header 2723, the TCP header 2724, and the data 2725 are copied to the IP layer. Next, the IP layer uses the I
A checksum value is obtained by using the P header 2723 as data to be added by the high-speed addition system 2712. Depending on the checksum value,
When it is determined that the received data is correct, when the processing of the IP layer is completed, the IP header 2723 is removed and the TCP header 27
Copy 24 and data 2725 to the TCP layer. If it is determined that the received data is incorrect, the received data is immediately discarded, and the processing on the receiving side ends. Next, T
The CP layer adds a pseudo header 2720, which is added only in the TCP layer to improve the accuracy of error determination of received data, before the TCP header 2724 of the copied data. Then, the TCP layer obtains the checksum value by using the pseudo header 2720, the TCP header 2724 and the data 2725 as the data to be added of the high speed addition system 2712. When it is determined that the received data is correct based on the checksum value, when the processing of the TCP layer is completed, the pseudo header 2720 and the TCP header 2724 are removed, the data 2725 is copied to the user application unit 2704, and the processing on the receiving side is performed. Ends. If it is determined that the received data is erroneous, the received data is immediately discarded and the processing on the receiving side ends.

As described above, by providing the high-speed addition system and the protocol buffer for storing the data to be added in the protocol control unit, the addition processing can be performed at high speed and the error data can be detected early.

Next, description will be made with reference to FIG. FIG. 28
Shows a configuration in the case where the network interface section includes the above-mentioned high-speed addition system and a network buffer for storing data to be added. In this embodiment, the protocol control unit transfers the processed information and the management data of the information to the network interface unit after the processing according to the protocol at the time of transmission, and receives the received information at the time of reception. The processing according to the protocol is performed, the network interface unit includes the addition unit, and the addition and reception process is performed based on the information transferred from the protocol control unit and the management data of the information at the time of transmission. To the section to be added data, add the checksum data in the adding section to the information and send it to the transmission path, and when receiving,
Receiving information from the transmission path, instructing the data to be added to the adder based on the received information, based on the checksum data in the adder, to determine whether the received data is correct, If it is correct, the process moves to the next predetermined process, and if it is not correct, the received data is discarded.

First, the processing on the transmitting side will be described.
The user application unit 2801 generates transmission data 2814 of 4430 bytes because the maximum data length permitted on the FDDI is 4.5 kilobytes. And
The transmission data 2814 is copied to the protocol buffer 2808 provided in the protocol control unit 2802, and the processing of the user application unit 2801 ends. Protocol control unit
In the 2802, first, the TCP layer adds a 20-byte pseudo header 2815 that is added only in the TCP layer to improve the accuracy of error detection of transfer data, and a 20-byte TCP header 2816 that stores TCP protocol information. Is added in front of the copied transmission data 2814, and the protocol buffer 28
Store in 08. And pseudo header 2815 and TCP header
2816 and data 2814 are copied to the IP layer, and the processing of the TCP layer ends. Next, the IP layer adds a 20-byte IP header 2817 storing IP protocol information to the pseudo header 28 before the TCP header 2816 of the copied data.
Add in parallel with 15. Then, the pseudo header 2815, the IP header 2817, the TCP header 2816, and the data 2814 are copied to the driver layer, and the processing in the IP layer ends. Next, the driver layer stores the 22-byte FDDI header 2818, which stores the physical address information, in the IP of the copied data.
Add before header 2817. And the FDDI header 28
18, IP header 2817, pseudo header 2815, TCP header 28
16 and data 2814 in the network interface unit 28
The data is copied to the network buffer 2809 provided in 03, and the processing of the driver layer ends. In the network interface section, first of all, the pseudo header 2815 and the TCP header 2815 among the data copied by the high-speed addition system 2810.
A checksum value is obtained by using 16 and data 2814 as data to be added, and the checksum value is assigned to the TCP header 2816. Next, the high-speed addition system 2810 uses the IP header 2817.
A checksum value is obtained as the data to be added, and the checksum value is assigned to the IP header 2817. Finally, the network interface unit 2803 has an FDDI header 2818 and an IP.
The header 2817, the TCP header 2816, and the data 2814 are sent to the LAN 2807 as a transmission frame 2819, and the processing on the receiving side ends.

Next, the processing on the receiving side will be described. First, the network interface unit 2806 changes the LAN 2807
FDDI header 2821 sent from the sending side via
A reception frame 2820 including an IP header 2822, a TCP header 2823, and data 2824 is received by the network buffer 2812. Next, the network interface unit 2806
In order to improve the accuracy of error determination of received data,
Pseudo header 2825 added only in P layer is IP header
It is created from 2822 and is added in parallel with the IP header 2822 before the TCP header 2823 of the received data. Then, the high speed addition system 2813 first obtains the checksum value using the IP header 2822 as the data to be added. When the checksum value determines that the received data is correct, the high speed addition system 2813 further determines that the pseudo header 2825 and the TCP header
The checksum value is obtained by using 2823 and the data 2724 as data to be added by the high-speed addition system 2813. If the checksum value determines that the received data is correct, the pseudo header 2825 is removed, and the FDDI header 2821, the IP header 2822, the TCP header 2823, and the data 2824 are added to the protocol control unit.
The data is copied to the protocol buffer 2811 provided in the 2805, and the processing of the network interface unit 2806 ends. If it is determined that the received data is erroneous, the received data is immediately discarded and the processing on the receiving side ends. In the protocol control unit 2805, the driver layer first processes the copied data, and when the processing is completed, F
The DDI header 2821 is removed, and the IP header 2822, the TCP header 2823, and the data 2824 are copied to the IP layer. Next,
The IP layer processes the copied data, removes the IP header 2822 when the processing is completed, and removes the TCP header 28
23 and the data 2824 are copied to the TCP layer, and the processing of the IP layer ends. Next, the TCP layer processes the copied data. When the processing is completed, the TCP header 2823 is removed, the data 2824 is copied to the user application unit 2804, and the processing on the receiving side is completed.

Next, a data management method performed in the above high speed communication system will be described with reference to FIGS. 28, 33 and 34. Conventionally, in the TCP layer, a pseudo header 2815 and a TCP header 2816 are created and added, and after the processing in the TCP layer is completed, the pseudo header 2815 is deleted, and the TCP header 2816 and the data 2814 are I.
It was passed to the P layer. In this high-speed communication system, since the high-speed addition system 2810 provided in the network interface unit 2803 performs checksum processing, up to the network buffer 2809 provided in the network interface unit 2803, a pseudo header that is one of the checksum processing targets. You have to pass 2825. Therefore, the data management table 3301 shown in FIG.
The data is managed by.

First, the structure of the data management table 3301 will be described with reference to FIG. The next table pointer 3302 stores a pointer that points to the next table when there are a plurality of data management tables, and 0 is entered when there is one table. The data pointer 3303 stores a pointer indicating the head position of the data managed by the table. The data length 3304 stores the size of data managed by the table. Unnecessary data length 3305
Among the data managed by the table, the size of the unnecessary data at the beginning which is not originally necessary in the lower layer, such as a pseudo header, is stored, and 0 is entered when there is no unnecessary data at the beginning.

Next, an example of using the data management table will be described with reference to FIG. In the TCP layer, first, the pseudo header 3406 and the TCP header 3407 are created and stored in the protocol buffer 2808, and then the data management table 3401 is created. Then, 0 is assigned to the next table pointer 3402, the address in which the pseudo header 3406 is stored is assigned to the data pointer 3403,
The data length 3404 includes the size of the pseudo header 3406,
A value obtained by adding the size of the TCP header 3407 and the size of the data 3408 is substituted, and the size of the pseudo header 3406 is substituted for the unnecessary data length 3405. After the processing in the TCP layer is completed, the data management table 3401 and the data are passed to the IP layer. Next, in the IP layer, first, the IP header 3414 is created and the data management table 3409 is created. Then, the address of the data management table 3401 passed from the TCP layer is assigned to the next table pointer 3410, and the IP header 3414 is assigned to the data pointer 3411.
Is stored, and the data length 3412
Is substituted with the size of the IP header 3414, and 0 is substituted for the unnecessary data length 3413. After the processing in the IP layer is completed, the data management table 340 created in the driver layer in the IP layer
9 and the data management table 3401 created in the TCP layer and the data. Next, in the driver layer, an FDDI header is created and added in front of the IP header 3414, the address of the FDDI header 3415 is assigned to the data pointer 3411, and the FDDI head 34 is assigned to the data length 3412.
A value obtained by summing the size of 15 and the size of the IP header 3414 is substituted. Since the FDDI header is a fixed type, it can be managed together with the IP header by the data management table 3409. After the processing in the driver layer is completed, the data management table 3409 created in the IP layer and the data management table 34 created in the TCP layer in the network interface section.
Pass 01 and the data. The network interface unit refers to the data management table 3409 created in the IP layer to refer to the FDDI header 3415, the IP header 3414, and the data management table 3 created in the TCP layer.
401 can be used, and the pseudo header 3406, TCP header 3407, and data 3408 can be used by referring to the management table 3401 created in the TCP layer. After the processing in the network interface section is completed, the FDDI header 3415, the IP header 3414, the TCP header 3407, and the data 3408, which are necessary for transmission, are transmitted by referring to the data management table.

As described above, in the high-speed communication system, it is possible to independently perform the addition process of adding the checksum data for the data by managing the data. When the checksum processing requests are transmitted one after another, the identification information may be added to the checksum processing requests.

As described above, by providing the network interface unit with the high-speed addition system and the network buffer for storing the data to be added, the addition processing can be performed at high speed and the detection of error data can be detected at an early stage. it can. Further, the header processing and the addition processing are independently processed, so that they can be processed in parallel.

Next, description will be made with reference to FIG. Figure 29
, The network interface unit has a high-speed addition system, the protocol control unit stores the data to be added,
The configuration in the case of having a common buffer shared by software and hardware is shown.
In the present embodiment, the addition unit is provided in the network interface unit, and the protocol control unit, at the time of transmission, instructs the addition unit to be added data and adds the checksum data in the addition unit to the information. Then, at the time of reception, the information is received from the transmission path, and the added data is instructed to the addition unit based on the received information, and the network interface unit receives the checksum data in the addition unit at the reception. Based on this, it is determined whether or not the received data is correct. If the received data is correct, the process proceeds to a predetermined next process, and if not, the received data is discarded. Further, the protocol control unit includes a common buffer that stores the information, and instructs the addition unit to specify an area of the added data and a storage area of the checksum data that are stored in the common buffer. The adder refers to the area of the data to be added instructed by the protocol control unit to perform addition, and stores the checksum data in the storage area of the instructed checksum data to convert the checksum data into information. In addition, the interface network unit instructs the common buffer to output transmission information at the time of transmission,
Upon reception, the received data can be stored in the common buffer.

First, the processing on the transmitting side will be described.
The user application unit 2901 generates the transmission data 2912 of 4430 bytes because the maximum data length permitted on the FDDI is 4.5 kilobytes. And
The transmission data 2912 is copied to the common buffer 2908 provided in the protocol control unit 2902, and the processing of the user application unit 2901 ends. Protocol control unit 2902
Then, first, the TCP layer includes a 20-byte pseudo header 2913 added only in the TCP layer in order to improve the accuracy of error detection of transfer data, and a 20-byte TCP header 2914 in which TCP protocol information is stored. , Is added before the copied transmission data 2912. Then, the high-speed addition system 29 provided in the network interface unit 2903
Sends a checksum processing request to 09. The checksum processing request includes a pseudo header 2913, a TCP header 2914, and data 2912 to be the checksum processing target.
The common buffer storage position, the data length, and the checksum processing result common buffer storage position are notified. Upon receiving the request, the high-speed addition system 2909 operates in parallel with the header processing in the protocol control unit 2902. The high-speed addition system 2909 refers to the common buffer 2908 from the notified storage position of the added data, obtains a checksum value using the pseudo header 2913, the TCP header 2914, and the data 2912 as the added data, and notifies the checksum The checksum value is stored in the processing result storage location. As a result, the checksum value can be substituted in the TCP header 2914. After storing the checksum value, high-speed addition system 2909
Notifies the protocol control unit 2902 of the end of the addition process. Then, when the TCP layer of the protocol control unit 2902 is notified of the addition processing, the pseudo header 2913 is removed, the TCP header 2914 and the data 2912 are copied to the IP layer, and the processing of the TCP layer ends. Next, the IP layer adds a 20-byte IP header 2915 in which IP protocol information is stored before the TCP header 2914 of the copied data.
Then, the checksum processing request is issued to the high-speed addition system 2909 provided in the network interface unit 2903. Similarly, the checksum processing request is made by notifying the common buffer storage position and length of the IP header 2915 to be the checksum processing target and the common buffer storage position of the checksum processing result. The high-speed addition system 2909 that received the request is the protocol control unit 2902.
It operates in parallel with the header processing in. The high-speed addition system 2909 refers to the common buffer 2908 from the notified storage position of the added data, obtains a checksum value using the IP header 2915 stored in the common buffer 2908 as the added data, and notifies the notified checksum. The checksum value is stored in the processing result storage location. This allows the checksum value to be assigned to the IP header 2915. After storing the checksum value, the high speed addition system 2909 notifies the protocol control unit 2902 of the end of the addition process. Then, in the IP layer of the protocol control unit 2902, the IP header 2915 and the T
The CP header 2914 and the data 2912 are copied to the driver layer, and the processing at the IP layer ends. Next, the driver layer is a 22-byte FD that stores physical address information.
The DI header 2916 is replaced with the IP header 29 of the copied data.
It is added before 15 and the transmission frame 29 is sent to the common buffer 2908.
Create 17. And the network interface section
The 2903 transmits the transmission frame 2917 from the common buffer 2908 to the L level.
The data is sent to AN2907, and the processing on the transmission side is completed.

When data is stored in the common buffer, when the data of each layer of the protocol control unit is delivered, the copy of the data includes the notification of the address indicating the position of the data stored in the common buffer. .
For example, when passing data from the TCP layer to the IP layer and from the IP layer to the driver layer, the address of the common buffer can be notified.

Next, the processing on the receiving side will be described.
First, the network interface unit receives, in the common buffer 2910, the reception frame 2918 including the FDDI header 2919, the IP header 2920, the TCP header 2921, and the data 2922 sent from the transmission side via the LAN 2907.

Next, the driver layer of the protocol control unit 2905 uses T in order to improve the accuracy of error determination of received data.
A pseudo header 2923 added only in the CP layer is created from the IP header 2920, and the TCP header 2921 of the received data is created.
Is added in parallel with the IP header 2920. Then, the checksum processing request is issued to the high-speed addition system 2911 provided in the network interface unit 2906. The checksum processing request is made by notifying the common buffer storage position and the length of the IP header 2920 and the pseudo header 2923 to be the checksum processing target, and the common buffer storage position of the checksum processing result. The high-speed addition system 2911 that received the request uses the protocol control unit 2905.
It operates in parallel with the header processing in. The high-speed addition system 2911 first refers to the common buffer 2910 from the storage position of the notified data to be added, and obtains the checksum value using the IP header 2920 stored in the common buffer 2910 as the data to be added. The checksum value is used to determine whether the received data is correct. When the received data is determined to be correct, the high-speed addition system 2911 further operates in parallel with the header processing in the protocol control unit 2905,
The checksum value is obtained by using the pseudo header 2923, the TCP header 2921, and the data 2922 as data to be added by the high-speed addition system 2911. The checksum value is used to determine whether the received data is correct. If it is determined that the received data is correct, the protocol control unit 2906 determines that the pseudo header 2923 and the FD
The DI header 2919 is removed, the IP header 2920, the TCP header 2921, and the data 2922 are copied to the IP layer, and the processing of the driver layer ends. If it is determined that the received data is erroneous, the received data is immediately discarded and the processing on the receiving side ends.

Next, the IP layer processes the copied data. When the processing is completed, the IP header 2920 is removed, the TCP header 2921 and the data 2922 are copied to the TCP layer, and the processing of the IP layer is completed. . Next, the TCP layer processes the copied data, and when the processing is completed, T
The CP header 2921 is removed, the data 2922 is copied to the user application unit 2904, and the processing on the receiving side ends.

As described above, the high-speed addition system and the network buffer for storing the data to be added are provided in the network interface section, and the common buffer is provided in the protocol control section. Data detection can be detected at an early stage. Further, the header processing and the addition processing are independently processed, so that they can be processed in parallel.

Next, description will be made with reference to FIG. Figure 30
Shows a configuration in the case where the network interface unit includes a high-speed addition system and a thumb buffer for temporarily storing the processing result of the high-speed addition system.
In this embodiment, when data is transferred between the protocol buffer provided in the protocol control unit 3002 and the network buffer provided in the network interface unit, the transfer data is extracted in the high-speed addition system to be added data. And In the present embodiment, the addition unit is provided in the network interface unit, and the addition unit receives transmission information from the protocol control unit at the time of transmission, performs addition processing based on the transmission information, and performs checksum data. And output to the network interface unit, and when receiving, add processing is performed based on the information received by the network interface unit to output checksum data, and transfer the received information to the protocol control unit. The network interface unit, when receiving, determines whether the received data is correct, based on the checksum data in the adding unit,
If it is correct, the process moves to the next predetermined process, and if it is not correct, the received data is discarded.

First, the processing on the transmitting side will be described. The user application unit 3001 determines that the maximum data length allowed on FDDI is 4.5 kilobytes.
Transmission data 3016 of 4430 bytes is generated. Then, the transmission data 3016 is copied to the protocol buffer 3008 provided in the protocol control unit 3002, and the processing of the user application unit 3001 ends. In the protocol control unit 3002, first, the TCP layer has a 20-byte pseudo header 3017 added only in the TCP layer to improve the accuracy of error detection of transfer data, and a 20-byte TCP in which TCP protocol information is stored. A header 3018 and a header 3018 are added before the copied transmission data 3016. Then, the pseudo header 3017, the TCP header 3018, and the data 3016 are copied to the IP layer, and the processing of the TCP layer ends. Next, the IP layer is a 20-byte I that stores IP protocol information.
The P header 3019 is replaced with the TCP header 30 of the copied data.
It is added in parallel with the pseudo header 3017 before 18. And
The pseudo header 3017, the IP header 3019, the TCP header 3018, and the data 3016 are copied to the driver layer, and the processing in the IP layer ends. Next, the driver layer adds a 22-byte FDDI header 3020 storing physical address information in front of the IP header 3019 of the copied data. After the processing in the driver layer is completed, the protocol control unit 3002 notifies the high-speed addition system 3011 that data will be output, and then sends the FDDI header 3020, the IP header 3019, the pseudo header 3017, the TCP header 3018, and the data 3016. Output.
The high-speed addition system 3011 first obtains a checksum value A3021 using the pseudo header 3017, the TCP header 3018, and the data 3016 as input data among the input data, and outputs the checksum value A3021 to the sum buffer 3010. Next, the checksum value B3022 is obtained using the IP header as the data to be added, and the checksum value B3022 is stored in the sum buffer 30.
Output to 10. After the addition processing is completed, the high-speed addition system 3011 removes the pseudo header 3017 from the data copied from the protocol control unit 3002 and outputs the FDDI header 3020, the IP header 3019, the TCP header 3018, and the data 3016 to the network buffer 3009. To do. The network interface unit 3003 has a checksum value A3021 and a checksum value B3022 stored in the sum buffer 3010.
Are substituted into the TCP header 3018 and the IP header 3019, respectively. Finally, the network interface unit 3003
Uses the FDDI header 3020, the IP header 3019, the TCP header 3018, and the data 3016 as the transmission frame 3023 on the LAN.
The data is sent to 3007, and the processing on the transmission side ends.

Next, the processing on the receiving side will be described.
First, the network interface unit 3006 uses the LAN 30
FDDI header 30 sent from the sender via 07
25, an IP header 3026, a TCP header 3027, and data 3028.
To receive. Next, the network interface unit 30
06 is T in order to improve the accuracy of error determination of received data.
A pseudo header 3029 added only in the CP layer is created from the IP header 3026, and the TCP header 3027 of the received data is created.
Is added in parallel with the IP header 3026. And F
High-speed addition system for DDI header 3025, IP header 3026, pseudo header 3029, TCP header 3027, and data 3028
Output to 3015. The high-speed addition system 3015 first obtains the checksum value A3030 using the IP header 3026 as the data to be added from the input data, and then checks the checksum value A30.
30 is output to the sum buffer 3014. Next, the pseudo header
3029 and TCP 3027 Header and data 3028 are used as augend data to obtain a checksum value B3031 and a checksum value B30
31 is output to the sum buffer 3014. After the addition processing is completed, the high speed addition system 3015 uses the data copied from the network interface unit 3006 to generate the pseudo header 3029.
, FDDI header 3025, IP header 3026 and TCP
The header 3027 and the data 3028 are output to the protocol buffer 3012. The network interface unit 3006 determines whether or not the received data is correct based on the checksum value A3030 and the checksum value B3031 stored in the sum buffer 3014. If it is determined to be correct, the protocol control unit 3005 continues processing, and if it is determined to be incorrect, the received data is immediately discarded, and the processing on the receiving side ends. In the protocol control unit 3005, when there is an input in the protocol buffer 3012, the driver layer first processes the copied data, and when the processing is completed, the FDDI header 3025 is removed and the IP header 3026 and the TCP header 3027 are stored. Copy data 3028 and IP layer. Next, the IP layer processes the copied data. When the processing is completed, the IP header 3026 is removed and the TC
Copy the P header 3027 and data 3028 to the TCP layer. Next, the TCP layer processes the copied data, removes the TCP header 3027 when the processing is completed, and
After copying 3028 to the user application unit 3004, the processing on the receiving side is completed.

By performing the above processing, when data is transferred between the protocol buffer and the network buffer, the transfer data is added as data to be added in the high-speed addition system via the high-speed addition system, and an error is generated in the network interface section. Detection can be performed.

Next, description will be made with reference to FIG. Figure 31
Shows a configuration in the case where the network interface unit includes a high-speed addition system and a thumb buffer. The present embodiment is characterized in that the data copied from the protocol buffer to the network buffer is the augend data at the time of transmission, and the data copied from the LAN to the network buffer is the augend data at the time of reception. . In this embodiment, the protocol control unit transfers the processed information and the management data of the information to the network interface unit after the processing according to the protocol at the time of transmission, and receives the information at the time of reception. Processing according to the above protocol, the addition unit is provided in the network interface unit,
The addition unit receives transmission information from the protocol control unit at the time of transmission, performs addition processing based on the transmission information, outputs checksum data and transmission information, and at the time of reception,
Information is received from the transmission path, addition processing is performed based on the information, checksum data is output, the information is output to the network interface unit, and the network interface unit outputs from the addition unit during transmission. Checksum data is added to the transmitted information that has been sent, and the transmission information is sent to the transmission path.At the time of reception, it is determined whether or not the received data is correct based on the checksum data in the adding unit. Shifts to the next predetermined process, and discards the received data if it is incorrect.

First, the processing on the transmitting side will be described. Since the maximum data length permitted on the FDDI is 4.5 kilobytes, the user application unit 3101
Transmission data 3116 of 4430 bytes is generated. Then, the transmission data 3116 is copied to the protocol buffer 3108 provided in the protocol control unit 3102, and the processing of the user application unit 3101 ends.

In the protocol control unit 3102, first, TCP
The layer is used by the TC to improve the accuracy of error detection of the transmitted data.
20-byte pseudo header 3117 added only in the P layer, and T
20-byte TCP that stores CP protocol information
A header 3118 is added before the copied transmission data 3116. Then, the pseudo header 3117, the TCP header 3118, and the data 3116 are copied to the IP layer, and the processing of the TCP layer ends. Next, the IP layer adds a 20-byte IP header 3119 in which IP protocol information is stored in parallel with the pseudo header 3117 before the TCP header 3118 of the copied data. Then, the pseudo header 3117 and the IP header 31
19, the TCP header 3118 and the data 3116 are copied to the driver layer, and the processing of the IP layer ends. Next, the driver layer is a 22-byte FD that stores physical address information.
The DI header 3120 is replaced with the IP header 31 of the copied data.
Add before 19. After the processing of the driver layer is completed, the protocol processing unit 3102 receives the FDDI header 3120, the IP header 3119, the pseudo header 3117, the TCP header 3118, and the data 3116.
It outputs to the high-speed addition system 3111 provided in the network interface unit. The high-speed addition system 3111
Of the input data, first the pseudo header 3117 and TC
A checksum value A3121 is obtained using the P header 3118 and the data 3116 as data to be added, and the checksum value A3121 is output to the sum buffer 3110. Next, the checksum value B3122 is obtained using the IP header as the data to be added, and the checksum value B3122 is output to the sum buffer 3110. After the addition processing is completed, the high-speed addition system 3111 removes the pseudo header 3117 from the data copied from the protocol control unit 3102 and outputs the FDDI header 3120, the IP header 3119, the TCP header 3118, and the data 3116 to the network buffer 3109. To do. The network interface unit 3103 transfers the checksum value A3121 and the checksum value B3122 stored in the sum buffer 3110 to the TCP header 3118 and the IP.
Substitute in header 3119 and respectively. Finally, the network interface unit 3103 sends the FDDI header 3120, the IP header 3119, the TCP header 3118, and the data 3116 as a transmission frame 3123 to the LAN 3107, and the processing on the transmission side ends.

Next, the processing on the receiving side will be described. First, the FDD sent from the sender via the LAN3107
A received frame 3124 including an I header 3125, an IP header 3126, a TCP header 3127 and data 3128 is input to the high speed addition system 3115. The fast addition system 3115
Of the input data, the IP header 3126 is used as the data to be added to obtain the checksum value A3130, and the checksum value A31
30 is output to the sum buffer 3114. Next, a checksum value B3131 is obtained using the TCP header 3127 and the data 3128 as data to be added, and the checksum value B3131 is output to the sum buffer 3114. After the addition processing is completed, the high-speed addition system 3115 receives the FDDI header, which is the input data.
3125, IP header 3126, TCP header 3127, and data 3128
And are output to the network buffer 3113. Next, the network interface unit 3106 creates the pseudo header 3129 added only in the TCP layer from the IP header 3126 in order to improve the accuracy of error determination of the received data, and outputs it to the high speed addition system 3115. The high-speed addition system 3115 obtains the checksum value using the pseudo header 3126 as the data to be added, and the checksum value B stored in the sum buffer.
Add checksum value to 3131. The network interface unit 3106 determines whether or not the received data is correct based on the checksum value A3130 and the checksum value B3131 stored in the sum buffer 3114. If it is determined that the received data is correct, the FDDI header 3125 and the IP header 3126 are used. And T
The protocol buffer 31 stores the CP header 3127 and the data 3128.
Copy to 12. If it is determined to be incorrect, the received data is immediately discarded, and the processing on the receiving side ends. In the protocol control unit 3105, the driver layer first processes the copied data, and when the processing is completed, the FD is processed.
Remove DI header 3125, IP header 3126 and TCP header
Copy 3127 and data 3128 to the IP layer. Next, IP
The layer processes the copied data, removes the IP header 3126 when the processing is completed, and copies the TCP header 3127 and the data 3128 to the TCP layer. Next, the TCP layer processes the copied data. When the processing is completed, the TCP header 3127 is removed, the data 3128 is copied to the user application unit 3104, and the processing on the receiving side is completed.

By the above processing, the data copied from the protocol buffer to the network buffer is used as the augend data at the time of transmission, and LA is received at the time of reception.
By using the data copied from N to the network buffer as the data to be added, and performing the addition process on the data to be added in the high-speed addition system, the error can be detected in the network interface unit.

Next, description will be made with reference to FIG. Figure 32
Shows a configuration diagram when the network interface unit includes a high-speed addition system and a thumb buffer, and the protocol control unit 3002 includes a common buffer. In the present embodiment, during transmission, the data in the common buffer of the protocol control unit 3002 is the augend data, and during reception,
It is characterized in that the data is transferred from the LAN to the common buffer via the high-speed addition system. In this embodiment, the adder is provided in the network interface, and the protocol controller instructs the adder to add data at the time of transmission and adds the checksum data in the adder to the information. When receiving, the adder receives information from the transmission path, performs addition processing based on the information, outputs checksum data, outputs the information to the protocol controller, and the network interface unit. At the time of reception, based on the checksum data in the addition unit, determines whether the received data is correct,
If it is correct, the process moves to the next predetermined process, and if it is not correct, the received data is discarded.

First, the processing on the transmitting side will be described.
The user application unit 3201 generates the transmission data 3213 of 4430 bytes because the maximum data length permitted on the FDDI is 4.5 kilobytes. And
The transmission data 3213 is copied to the common buffer 3208 provided in the protocol control unit 3202, and the processing of the user application unit 3201 ends. Protocol control unit 3202
Then, first, the TCP layer includes a 20-byte pseudo header 3214 that is added only in the TCP layer in order to improve the accuracy of error detection of transfer data, and a 20-byte TCP header 3215 that stores TCP protocol information. , Is added before the copied transmission data 3213. Then, the high-speed addition system 32 provided in the network interface unit 3203
Sends a checksum processing request to 09. The checksum processing request includes a pseudo header 3214, a TCP header 3215, and data 3213 that are the target of the checksum processing.
The common buffer storage position, the data length, and the checksum processing result common buffer storage position are notified. Upon receiving the request, the high speed addition system 3209 operates in parallel with the header processing in the protocol control unit 3202. The high-speed addition system 3209 uses the pseudo header 3214, the TCP header 3215, and the data stored in the protocol buffer 3208.
3213 and 3213 are used as augend data to obtain a checksum value, and the checksum value is substituted into the notified checksum value storage area. Then, the pseudo header 3214 is removed, and the TCP header
The 3215 and the data 3213 are copied to the IP layer, and the processing of the TCP layer ends. Next, the IP layer adds a 20-byte IP header 3216 in which IP protocol information is stored before the TCP header 3215 of the copied data. Then, the checksum processing request is issued to the high-speed addition system 3209 provided in the network interface unit 3203. Similarly, as a checksum processing request,
The common buffer storage position and length of the IP header 3216 to be subjected to the checksum processing and the common buffer storage position of the checksum processing result are notified and performed. Upon receiving the request, the high-speed addition system 3209 operates in parallel with the header processing in the protocol control unit 3202. High speed addition system 32
In 09, the checksum value is obtained using the IP header 3216 stored in the protocol buffer 3208 as data to be added,
The checksum value is assigned to the notified checksum value storage area. Then, the IP header 3216 and the TCP header 3215
And the data 3213 are copied to the driver layer, and the processing of the IP layer ends. Next, the driver layer adds a 22-byte FDDI header 3217 storing physical address information to the front of the IP header 3216 of the copied data to create a transmission frame 3218. After finishing the process in the driver layer,
The network interface unit 3203 uses the transmission frame 32.
18 is sent to the LAN 3207, and the processing on the transmission side ends.

Next, the processing on the receiving side will be described. First, the FD sent from the sending side via LAN3207
A received frame 3219 including a DI header 3220, an IP header 3221, a TCP header 3222, and data 3223 is input to the high-speed addition system 3212. The high-speed addition system 3212 first obtains a checksum value A3225 with the IP header 3221 of the input data as the data to be added, and outputs the checksum value A3225 to the sum buffer 3211. Next, TCP
A checksum value B3226 is obtained using the header 3222 and the data 3223 as data to be added, and the checksum value B3226 is output to the sum buffer 3211. After the addition processing is completed, the high-speed addition system 3212 receives the FDDI header 3220, the IP header 3221, the TCP header 3222, and the data 32 which are the input data.
23 and 23 are output to the protocol buffer 3211. Next, the protocol control unit 3205 creates the pseudo header 3224 added only in the TCP layer from the IP header 3221 in order to improve the accuracy of error determination of the received data, and the high speed addition system.
Output to 3212. High speed addition system 3212, pseudo header
A checksum value is obtained using 3224 as the data to be added, and the checksum value is added to the checksum value B3226 stored in the sum buffer. The network interface unit determines whether or not the received data is correct based on the checksum value A3225 and checksum value B3226 stored in the sum buffer 3211. If it is determined to be correct,
If the protocol control unit 3205 continues processing and it is determined that the data is incorrect, the received data is immediately discarded,
The processing on the receiving side ends. The driver layer processes the copied data, and when the processing is completed, the FDDI
Remove the header 3220, IP header 3221 and TCP header 3222
And data 3223 are copied to the IP layer. Next, the IP layer processes the copied data, removes the IP header 3221 when the processing is completed, and removes the TCP header 3222 and the data 32.
Copy 23 to TCP layer. Next, the TCP layer processes the copied data, and when the processing is completed, TC is executed.
The P header 3222 is removed, the data 3223 is copied to the user application section 3204, and the processing on the receiving side ends.

By the above processing, the data in the common buffer is used as the data to be added during transmission, and the data transferred from the LAN to the common buffer via the high-speed addition system is used as the data to be added during reception. The error can be detected in the network interface unit by adding data to be added in the high-speed addition system.

[0138]

According to the present invention, high-speed processing can be performed in the addition system. In addition, communication processing can be performed at high speed in a communication system including an addition system.

[Brief description of drawings]

FIG. 1 is a configuration diagram when an addition system using 1's complement addition is used for error detection processing of a network communication system.

FIG. 2 is a flowchart showing a processing flow of a conventional addition system.

FIG. 3 The conventional addition system performs 4470 bytes of added data stored from an even address 2
Explanatory drawing which shows a byte length addition process.

FIG. 4 is an explanatory diagram showing a specific example of a 2-byte length addition process performed by a conventional addition system on 16-byte added data stored from an even address.

FIG. 5: 1 performed by a conventional addition system on 4470-byte added data stored from an odd address
Explanatory drawing which shows a byte length addition process.

FIG. 6 is an explanatory diagram showing one's complement addition for two-digit augend data used in the addition system of the present invention.

FIG. 7 is an explanatory diagram showing an example in which the present invention is used for checksum processing on the transmitting side of a network communication system using the TCP / IP protocol.

FIG. 8 is an explanatory diagram showing an example in which the present invention is used for checksum processing on the receiving side of a network communication system based on the TCP / IP protocol.

FIG. 9 is an explanatory diagram showing one's complement addition for n-digit added data used in the addition system of the present invention.

10 is an explanatory diagram showing one's complement addition to the augend data in which the upper digit and the lower digit of the augend data of FIG. 6 are used, which is used in the addition system targeted by the present invention.

FIG. 11 is a flowchart showing a processing flow of the present invention.

FIG. 12 is a block diagram showing the configuration of a high-speed addition system.

FIG. 13 is a hardware configuration diagram of addition processing using two-step 1's complement addition.

FIG. 14 is an explanatory diagram showing a program of an addition process using a one's complement addition process in C language.

FIG. 15 is an explanatory diagram showing a program of addition processing using a one's complement addition processing in assembler language.

FIG. 16 is an explanatory diagram of 4-byte length addition performed on 4470-byte added data that is suitable for a boundary according to the present invention.

FIG. 17 is an explanatory diagram of 4-byte length addition performed on 16-byte added data of a specific numerical value suitable for a boundary according to the present invention.

FIG. 18 is a 4-byte operation to be performed on 4470-byte added data, which is unsuitable for the boundary because the present invention is stored from odd addresses and there is a boundary between the first 1st byte and the 2nd byte. Explanatory drawing of long addition.

FIG. 19 is a specific numerical value of 1 according to the present invention, which is unsuitable for a boundary because it is stored from an odd address and there is a boundary between the first byte and the second byte.
Explanatory drawing of 4-byte length addition performed for 6-byte added data.

FIG. 20 is a 4-byte operation performed on 4470-byte added data according to the present invention, which is unsuitable for a boundary because it is stored from an even-numbered address and there is a boundary between the first and second bytes. Explanatory drawing of long addition.

FIG. 21 is a specific numerical value of 16 according to the present invention, which is unsuitable for a boundary because it is stored from an even-numbered address and there is a boundary between the first and second bytes.
Explanatory drawing of the 4-byte length addition process performed with respect to the data to be added of bytes.

FIG. 22 is a diagram showing a case where the present invention is unsuitable for the boundary because the boundary is stored between the 3rd and 4th bytes at the beginning, which is stored from an odd address, and is performed for 4470 bytes of added data. Explanatory drawing of a long addition process.

FIG. 23 is a specific numerical value of 16 according to the present invention, which is unsuitable for a boundary because it is stored from an odd address and there is a boundary between the 3rd and 4th bytes at the beginning.
Explanatory drawing of 4-byte length addition performed for data to be added of bytes.

FIG. 24 is a hardware configuration diagram of an addition process using one's complement addition in one step.

FIG. 25 is an explanatory diagram for explaining an example of adding 4-byte length.

FIG. 26 is an explanatory diagram of a case where positions are exchanged and addition is performed.

FIG. 27 is a diagram showing processing when the protocol control unit is provided with a high-speed addition system and data in the protocol buffer is used as data to be added;

FIG. 28 is a diagram showing processing when a high-speed addition system is provided in the network interface unit and data in a network buffer is used as data to be added.

FIG. 29 is a diagram showing processing when a high-speed addition system is provided in the network interface unit and data in a common buffer is used as data to be added.

FIG. 30 is a diagram showing processing in the case where a high-speed addition system is provided in the network interface unit and data copied between a protocol buffer and a network buffer is used as data to be added.

FIG. 31 is a diagram showing a case where a high-speed addition system is provided in the network interface unit, and data to be added is data copied from the protocol buffer to the network buffer during transmission and data to be copied from LAN to the network buffer during reception. The figure which shows the process in case of making it addition data.

FIG. 32 shows a case where a high-speed addition system is provided in the network interface unit and data in the common buffer is used as data to be added during transmission and data copied from the LAN to the common buffer is used as data to be added during reception. The figure which shows a process.

FIG. 33 is a data management table.

FIG. 34 is an explanatory diagram showing an example of using the data management table.

[Explanation of symbols]

1101 ... Boundary determination processing of top data, 1102
Data saving process, 1103 ... 2 byte length compression process of saved data, 1104 ... 4 byte length addition process of added data, 1105 ... 2 byte length compression process of 4 byte length data, 1106 ... Boundary determination process of top data, 11
07 ... head data address determination processing, 1108 ... 2-byte length data correction processing, 1109 ... save data addition processing, 1110 ... addition result storage processing, 1201 ... added data storage area, 1202 ... head data determination unit,
1203 ... Data save section, 1204 ... Save data storage area, 1205 ... 2 byte length compression section, 1206 ... Compressed save data storage area, 1207 ... 4 byte length adder section, 1
208 ... 4-byte length data storage area, 1209 ... 2-byte length data storage area, 1210 ... Data correction unit, 1
211 ... Saved data addition unit, 1212 ... Addition result storage area.

 ─────────────────────────────────────────────────── ─── Continued front page (72) Inventor Yoshinori Watanabe 292 Yoshida-cho, Totsuka-ku, Yokohama-shi, Kanagawa Hitachi, Ltd. Microelectronics Device Development Laboratory

Claims (15)

[Claims] 1. An addition system for reading out data stored in a storage means, performing 1's complement addition, and outputting an addition result of a predetermined data length, the unit being larger than the predetermined data length. At 1 to read data from the storage means and sequentially perform 1's complement addition on the read data, and a compression to perform 1's complement addition by dividing the result of the addition unit into units of the predetermined data length. And an output unit that outputs a compression addition result of the compression unit. 2. The head data determining unit for determining whether or not the head portion of the data stored in the storage unit is the head unit of the reading unit of the storage unit, and the head data determining unit. As a result of the determination, when the read unit of the storage means is not the start portion, the data holding the data from the start portion of the data stored in the storage means to the break of the start portion of the first read unit of the storage means Further comprising a save unit, wherein the addition unit reads the data after the break from the storage unit in units of reading by the storage unit,
The read data is sequentially subjected to 1's complement addition, and the compression unit performs 1's complement addition by dividing the addition result of the data after the break of the addition unit into units of the predetermined data length. As a result of the judgment by the head data judgment unit, if the data is not the head part of the reading unit of the storage means, the data held in the data saving unit is divided into units of the predetermined data length and 1's complement addition is performed. Further, a save data addition unit that adds one's complement to the addition result in the compression unit for the data held in the data saving unit is further added to the compression addition result in the compression unit for the data after the break. An addition system characterized by having. 3. The addition according to claim 2, wherein when the predetermined data length is 2 bytes, the storage unit sets the read unit of the storage unit to an even number of bytes of 4 bytes or more. system. 4. The addition system according to claim 3, wherein the storage unit sets the read unit of the storage unit to 4 bytes. 5. The head data determination unit according to claim 4, further determining whether the storage address of the head portion of the data stored in the storage unit is an odd number or an even number, and the result of the determination by the head data determination unit. , A data correction unit for exchanging data between the upper byte and the lower byte of the compression addition result obtained by compressing the addition result of the data after the break in the addition unit in the addition unit when stored in an odd address An addition system, further comprising: 6. A communication control unit for controlling transmission / reception of information via a transmission line, a processing unit for processing information, and error detection data for correctly transmitting / receiving transmission information when transmitting / receiving information should be transmitted. An information processing device having a checksum unit that is added to information and detects an error from received information, wherein the checksum unit is a storage unit that holds the information, and information that is stored in the storage unit. Is provided with addition means for performing 1's complement addition by 4 bytes each, and compression means for performing 1's complement addition by 2 bytes for each addition result, and the addition result of the compression section is used as the error detection data. A characteristic information processing device. 7. The addition means according to claim 6, wherein the addition means determines whether or not the head portion of the information held in the storage means is the head portion of the reading unit of the storage means, and if it is not the head portion. Is a two-byte one-complement addition for the information from the head portion of the information held in the storage means to the break of the head portion of the first read unit of the storage means, and the information after the break is stored in the memory. 1-complement addition by 4 bytes from the means, and the addition result for the information after the break is further 1-complement addition by 2 bytes for compression, and the compression addition result for the information after the break; An information processing device, which adds and outputs an addition result of information up to a break. 8. The adder according to claim 7, further determining whether the storage address of the head portion of the information held in said storage means is an odd number or an even number. Data is exchanged between the upper byte and the lower byte of the compression addition result for the information after the break, and the value after the data exchange and the addition result for the information from the leading part to the break are added and output. A characteristic information processing device. 9. A communication system having a transmission line and a plurality of transmission / reception devices for transmitting / receiving information via the transmission line, wherein the transmission / reception device includes a network interface unit for performing transmission / reception processing on the transmission line, and transmission / reception information. A user application unit that processes the data, a protocol control unit that performs processing according to a predetermined protocol at the time of transmission / reception, and at the time of transmitting / receiving information, for adding / detecting checksum data for error detection for correctly transmitting / receiving transmission information. And an adder for adding data to be added. The adder divides the addition result into a unit of a predetermined data length for the one-complement addition. Compression means for performing one's complement addition, and the addition result of the compression section is checked by the error detection checker. Communication system, characterized in that the sum data. 10. The protocol control unit according to claim 9, further comprising the addition unit, which instructs the addition unit to perform addition-added data along with the processing according to the protocol, and the addition unit is instructed. And outputs checksum data by performing addition processing on the data to be added, and when receiving, the protocol control unit determines whether or not the received data is correct based on the checksum data in the adding unit. Is a communication system characterized by shifting to the next predetermined process and discarding the received data when it is incorrect. 11. The protocol control unit according to claim 9, at the time of transmission, after the processing according to the protocol is completed, the processed information and management data of the information are transferred to the network interface unit, and at the time of reception. Processing the received information in accordance with the protocol, the network interface unit includes the adding unit, and at the same time as the transmission / reception process, the information transferred from the protocol control unit at the time of transmission and management data of the information. Based on the above, the data to be added is instructed to the adder, the checksum data in the adder is added to the information, and the information is sent to the transmission line. Upon reception, the information is received from the transmission line and received. Based on the information, the adder is instructed which data to be added, and based on the checksum data in the adder, the received data is received. A communication system characterized by judging whether or not the received data is correct, and if the received data is correct, shifts to the next predetermined process, and if not, discards the received data. 12. The network interface unit according to claim 9, wherein the network interface unit includes the adder unit, and the protocol control unit instructs the adder unit to perform data addition when transmitting, and a checksum in the adder unit. Data is added to the information, the information is received from the transmission path at the time of reception, the data to be added is instructed to the addition unit based on the received information, and the network interface unit is provided at the time of reception Based on the checksum data in, it is determined whether or not the received data is correct. If it is correct, the process proceeds to the next predetermined process, and if it is not correct,
A communication system characterized by discarding received data. 13. The protocol control unit according to claim 12, further comprising a common buffer for storing the information, wherein an area of added data stored in the common buffer and the checksum data are stored in the addition unit. Storage area of the checksum data is stored in the storage area of the instructed checksum data, and the addition section refers to the area of the data to be added instructed by the protocol control section. According to the above, the checksum data is added to the information, and the interface network unit instructs the common buffer to output the transmission information at the time of transmission, and at the time of reception,
A communication system, wherein received data is stored in the common buffer. 13. The network interface unit according to claim 9, wherein the addition unit includes the addition unit.
At the time of transmission, the transmission information from the protocol control unit is received, addition processing is performed based on the transmission information, checksum data is added and output to the network interface unit, and at the time of reception, it is received by the network interface unit. The checksum data is output by performing addition processing based on the received information, the received information is transferred to the protocol control unit, and the network interface unit receives the received data based on the checksum data in the addition unit when receiving. It is judged whether it is correct, and if it is correct, it moves to the next predetermined process, and if it is not correct,
A communication system characterized by discarding received data. 14. The protocol control unit according to claim 9, at the time of transmission, after processing according to the protocol is completed, the processed information and management data of the information are transferred to the network interface unit, and at the time of reception. Processing the received information according to the protocol, the network interface unit includes the addition unit, the addition unit receives the transmission information from the protocol control unit at the time of transmission, and adds the transmission information to the transmission information. Based on the addition processing, output checksum data and transmission information, and when receiving,
Information is received from the transmission path, addition processing is performed based on the information, checksum data is output, the information is output to the network interface unit, and the network interface unit outputs from the addition unit when transmitting. Checksum data is added to the transmitted information that has been sent, and the transmission information is sent to the transmission path.At the time of reception, it is determined whether or not the received data is correct based on the checksum data in the adding unit. Is a communication system characterized by shifting to the next predetermined process and discarding the received data when it is incorrect. 15. The network interface unit according to claim 9, further comprising: the addition unit, wherein the protocol control unit, when transmitting, instructs the addition unit to perform data addition, and a checksum in the addition unit. The data is added to the information, and the addition unit receives the information from the transmission path at the time of reception, performs addition processing based on the information, outputs checksum data, and outputs the information to the protocol control unit. At the time of reception, the network interface unit determines whether or not the received data is correct based on the checksum data in the adding unit, and if correct, shifts to the next predetermined process Has
A communication system characterized by discarding received data.