JP3347362B2 - SCSI data transfer method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to a SCSI (Small Comp
The present invention relates to a SCSI data transfer method for exchanging information between an initiator and a target based on a uter system interface (standard).

[0002]

2. Description of the Related Art SCSI is an interface SA for connecting a small computer manufactured by Sugart Corporation in the United States and peripheral devices.
A based on SI (Shugart Associates System Interface)
It was discussed at the NSI (American Standards Association) X3T9.2 Working Committee and standardized as ANSI X3.131-1986.
It is currently becoming the standard interface between personal computers and their peripherals. Recently, work has been underway to standardize an extended version of SCSI-2, but a final decision has not been reached at this time. The range of the SCSI interface standard defined by the ANSI is the following five points. (1) Types of interface signals, their definitions, and signal transfer timings. (2) Definition of a protocol, each phase, and the like that define an operation sequence as an interface. (3) Physical interface conditions such as cable specifications and connector specifications, and electrical conditions of the transmission system. (4) A command system for executing various controls and data transfer of peripheral devices, a command format and a function of each command. (5) A status byte format for notifying the host computer of the execution result of the command, and a structure of sense data for notifying an abnormal state in the process.

The command system in the above item (4) is as follows. First, SCSI commands are divided into eight types of groups. CDB (Command Descript
The first byte of (er Block) is an operation code. The upper three bits specify a group code, and the lower five bits specify a command code (code indicating a command type) for each group.

[0004] The length of the CDB is defined for each group as follows. (1) Group 0: 6 bytes (2) Group 1: 10 bytes (3) Groups 2 to 4 Reserved (4) Group 5: 12 Byte (5) Groups 6 to 7... Vendor Unique (manufacturer's rule) The CDBs in Groups 6 to 7 are command groups that can be defined uniquely for SCSI devices. In each command,
In the logical block address, fixed-length data blocks are continuously arranged on the logical unit.

FIG . 6 is a diagram showing a configuration of a logical block in a hard disk device. In FIG. 6 , the data block of cylinder = 0 and sector = 0 is set to logical block address = 0, and the logical block address is set to 1 each time the data block is incremented by one in the order of sector, track and cylinder.
One. The logical block address is excellent in that the initiator (host computer) accesses the data by designating the logical block address of the first data block and the number of processing blocks, so that there is no need to be aware of the physical structure. . Therefore, by using the logical block address, when devices having different numbers of cylinders, tracks, sectors, and the like are connected to each other, it is possible to operate with the same software.

FIG . 7 is a diagram showing a general example of a SCSI system configuration. As a logical unit, a physical device such as a hard disk is often connected as shown in FIG. 7 , but a logical unit number (LUN: Logicala) is usually used .
l Unit Number) may be assigned to a physical device or may be assigned to a virtual device. Normal SC
In SI, LUN = 0 to 7 of 8 on the SCSI bus
Logical units can be connected, and a maximum of 2048 logical units can be connected by using extended messages.

The SCSI data line is -DB0
Parallel communication is performed by eight lines (from "-" indicating negative logic) to -DB7, but a dedicated parity bit line called -DBP is provided separately to secure data reliability. If the initiator detects a parity error in the data received from the target, the initiator
r If a parity error is detected in the message received from the target
Send an ssage Parity Error message to each target.

[0008] If the target supports these messages and has a retry function,
The target can recover from the error by sending data or a message to the initiator again. On the other hand, if the target detects a parity error in the command or data received from the initiator, if the target has a retry function,
After transmitting the store Pointers message to the initiator, returning to the previous phase (command phase or data phase), and re-receiving the command or data, the error can be recovered.

[0009]

In the related art, there are various causes of the parity error. One of the causes is that some noise such as power supply noise or external noise is mixed in the SCSI transfer line. In the case of a parity error due to such a cause, the error occurs only while noise is mixed, so that the error can be recovered by the method described above.

However, even when an abnormality occurs in a communication line such as a data line disconnection, a parity error may occur as a result depending on the manner of occurrence of the abnormality. The parity error thus generated continuously occurs unless the cause of the abnormality itself is removed, for example, the disconnection is restored. Therefore, it is impossible to recover the error by the above-described method. Also,
In addition, there is a problem in that the initiator and the target mutually exchange information indicating that a parity error has been detected, and retry without recovery is infinitely repeated.

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and an object of the present invention is to make it possible to recover an error even if a parity error occurs continuously unless the cause of the abnormality is removed. It is another object of the present invention to provide a SCSI data transfer method.

[0012]

SUMMARY OF THE INVENTION An object of the present invention is to provide a SCSI data transfer system for exchanging information between an initiator and a target based on the Small Computer System Interface (hereinafter abbreviated as SCSI) standard. During the processing of the command, during the retry process started by the target detecting a parity error, or during the retry process started by the initiator sending a report indicating that the parity error was detected to the target, the parity When an error is detected or a report of the detection is received, a check condition status is transmitted to the initiator, and when the check condition status cannot be transmitted normally, a command complete message is transmitted. Not It is achieved by the SCSI data transfer method which is characterized in that the process proceeds to scan-free phase.

[0013]

Embodiments of the present invention will be described below in detail with reference to the drawings. Here, the SCS of the present invention is used.
As an explanation of the embodiment of the I-data transfer method, based on the flowcharts of FIG. 1 to FIG. 5, when a parity error is detected during SCSI command processing or when a report indicating that a parity error is detected is received, Is performed. In addition, the SCSI command processing referred to here starts with arbitration and uses Command Comp
It refers to a series of operations that end with a lete message, not just the operation of the command phase.

FIG. 1 is a flowchart for explaining the processing procedure of the first embodiment of the present invention, and shows the flow of processing when a parity error is detected in data received by the target in the data out phase. It is. Less than,
The processing procedure will be described using a flowchart,
In the flowcharts and the description, the description of the phase transition will be omitted unless particularly necessary. For example, Check Co
In order to transmit the ndition status, the phase must be shifted to the status phase, but in the following, it is simply described as “transmit the Check Condition status”, and the process of shifting to the status phase is omitted. And

When a parity error is detected in the data received by the target in the data out phase, the following processing is performed. In FIG. 1, first, in step 101, the target sends a Restore Pointers message. next,
In step 102, it is checked whether the transmission is normal. If the transmission is successful, the process proceeds to the data out phase in step 103. Then, the initiator resends the data, and receives the data,
At step 104, it is checked whether or not the reception was successful. If the data is successfully received in step 104, it means that the retry processing has succeeded, so the retry processing is terminated and the next processing is performed.

On the other hand, if the data cannot be normally received in step 104, the flow advances to step 105 to execute the following processing. The reason why data cannot be received normally is that a parity error is detected again in the received data, that the initiator has transmitted an Initiator Detected Error message, that the data could not be transmitted normally in step 102, that is, that the transmitted Restore Pointers When a Message Parity Error message is received for a message, some message is transmitted, but there is a case where a parity error is detected and the message cannot be normally received.

In such a case, since a parity error has occurred continuously, it is determined that the error is fatal, and the sense data is set in step 105, and step 10 is executed.
6. Send Check Condition status. In Step 108, a Command Complete message is transmitted.
In step 109, the process shifts to the bus free phase and the process ends. In step 105, the sense key is Hardware Error and the sense code is SCSI Parity Er
Set ror.

Here, in the situation where the parity error continuously occurs as described above, the Chec
As described above, it is very unlikely that the k Condition status will be received by the initiator. Therefore, in step 107, it is checked whether or not the transmission has been normally performed (received by the initiator). If an error is detected and the message cannot be received normally, the process immediately shifts to the bus-free phase in step 109 without transmitting a Command Complete message, and ends the processing.

FIG. 2 is a flowchart for explaining the processing procedure of the second embodiment of the present invention, showing the flow of processing when a parity error is detected in a command received by the target in the command phase. is there. In this embodiment, when a parity error is detected in the command received by the target in the command phase, the following processing is performed. In FIG. 2, first, in step 201, the target sends a Restore Pointers message.

Next, at step 202, it is checked whether the transmission is normal. If the transmission is successful, the process proceeds to step 203 in the command phase. Then, since the initiator resends the command, the initiator receives the command and checks in step 204 whether or not the command was successfully received. If the data is successfully received in step 204, it means that the retry processing has succeeded, so the retry processing is terminated and the next processing is performed.

On the other hand, if the data cannot be normally received in step 204, the flow advances to step 205 to execute the next processing. The reason why the data cannot be received normally is that, if a parity error is detected again in the received command, if the initiator has transmitted an Initiator Detected Error message, if the transmission was not successful in step 202, Restore Pointers
If a Message Parity Error message is received for the message, some message has been sent,
There are cases where a parity error is detected and the message cannot be received normally.

In such a case, since a parity error has occurred continuously, it is determined that the error is fatal, and sense data is set in step 205 and
6. Send Check Condition status. In step 208, a Command Complete message is transmitted,
In step 209, the process shifts to the bus free phase and the process ends. In step 205, the sense key is Hardware Error and the sense code is SCSI Parity Err.
Set ror.

Even in this embodiment, if a parity error continues to occur, the Ch transmitted at step 206
Since the possibility that the eck Condition status is received by the initiator is very low as described above, it is checked in step 207 whether or not the eck condition status has been normally transmitted (received by the initiator). Then, it is not received normally, that is, the initiator
If you have sent a tected Error message or some kind of message, but have detected a parity error and have not received the message normally,
Without transmitting the and Complete message, the process immediately shifts to the bus free phase in step 209 and ends the process. Although not shown in FIG. 2, if the sense data cannot be set correctly due to, for example, an unknown LUN in step 205, the process immediately shifts to the bus-free phase in step 209 and ends the processing.

FIG. 3 is a flowchart for explaining a processing procedure according to a third embodiment of the present invention.
8 shows a flow of processing when a parity error is detected in a message received by the target in the phase. In this embodiment, when a parity error is detected in the message received by the target in the message out phase, the following processing is performed. In FIG. 3, first, at step 301, the process proceeds to the message out phase. Then, the initiator resends the message, so that the initiator receives the message and checks in step 302 whether the message was successfully received. Step 30
If the data is successfully received in step 2, the retry processing is successful, and the retry processing is terminated and the next processing is performed.

On the other hand, if the message cannot be received normally in step 302, that is, if a parity error is detected again in the received message, or if a message has been transmitted, the parity error is detected and the message cannot be received normally. If there is no such error, it means that a parity error has occurred continuously, so that it is determined that the error is fatal. Then, in step 303, sense data is set, and in step 304, a Check Condition status is transmitted. In step 306, a Command Complete message is transmitted, and in step 307, the process shifts to a bus-free phase and ends. In step 303, the sense key is Hardware Error and the sense code is SC.
Set SI Parity Error.

Also in this embodiment, if a parity error occurs continuously, the
Since the possibility that the Check Condition status is received by the initiator is very low, it is checked in step 305 whether or not the status has been normally transmitted (received by the initiator). If the message is not received normally, that is, the initiator has transmitted an Initiator Detected Error message or has transmitted some message, but has detected a parity error and has not been able to receive the message normally, it transmits a Command Complete message. Instead, the process immediately shifts to the bus free phase in step 307, and the process ends.

The message (Identify, Abort) received by the target during the selection with attention.
, Or a Bus Device Reset message), the same processing as the processing procedure of FIG. 3 is performed. However, in this case, for example, when the message transmitted by the initiator is an Identify message, the LU
Since N is unknown and sense data cannot be set correctly in step 303, the process immediately shifts to the bus free phase in step 307 and ends the processing.

The case where a parity error has been detected in the information (data, command, message) received by the target has been described above. Next, a message indicating that a parity error has been detected in the information received by the initiator is sent to the target. Will be described.

FIG. 4 is a flowchart for explaining the processing procedure of the fourth embodiment of the present invention, and shows the flow of processing when the target transmits an Initiator Detected Error message from the initiator. In this embodiment, the target is set to Initiator Detect from the initiator.
When the ed Error message is received, the following processing is performed. In FIG. 4, first, in step 401, the target sends a Restore Pointers message.

Next, at step 402, it is checked whether the transmission is normal. If the transmission is successful, the process returns to the phase immediately before receiving the Initiator Detected Error message in step 403 and returns to the Initiator Detected Error message.
or Retry the process that was being performed when the message was received. Next, at step 404, it is checked whether the retry process has been executed normally. If the retry processing has succeeded in step 404, it means that the retry processing has succeeded, so the retry processing ends and the next processing is performed.

On the other hand, if the transmission cannot be performed normally in step 404 and if the transmission cannot be performed normally in step 402, that is, if the initiator re-starts the Initiator Dete
If a ctedError message has been sent, or if a message has been sent, but a parity error has been detected and the message has not been successfully received, a fatal error has occurred because the parity error has occurred continuously. In step 405, the sense data is set, and in step 406, the Check Condition status is transmitted.

In step 408, Command Complete
Send a message, and in step 409,
Move to the phase and end the process. Step 40
In step 5, the sense key is Hardware Error, and the sense code is SCS
If the retry processing was not successful in step 404, the Initiator Detected Error Mes
Set sage Receives respectively.

Even in this embodiment, if a parity error continues to occur, the Ch transmitted at step 406
Since it is very unlikely that the eck Condition status will be received by the initiator, it is checked in step 407 whether the eck Condition status has been normally transmitted (received by the initiator). If the message is not received normally, that is, the initiator has transmitted an Initiator Detected Error message or has transmitted some message, but has detected a parity error and has not been able to receive the message normally, a Command Complete message is transmitted. Instead, the process immediately shifts to the bus free phase in step 409, and the process ends.

FIG. 5 is a flowchart for explaining the processing procedure of the fifth embodiment of the present invention, and shows the flow of processing when the target receives a Message Parity Error message from the initiator. In this example, the target is the Message Parity Err from the initiator.
When the or message is received, the following processing is performed.
In FIG. 5, first, in step 501, the message
The process proceeds to the phase, and the message is retransmitted. In step 502, it is checked whether the message has been successfully transmitted. If the transmission was successful in step 502, it means that the retry process was successful, and the retry process is terminated, and the next process is performed.

On the other hand, if the transmission is not successful in step 502, that is, if the initiator returns to Message Pari
If a ty Error message has been sent, or if a message has been sent, but a parity error has been detected and the message has not been received normally, a fatal error has occurred consecutively. When it is determined that an error has occurred, sense data is set in step 503, and a check condition status is transmitted in step 504. In step 506, a Command Complete message is transmitted, and in step 507, the process shifts to a bus free phase and ends. Step 50
At 3, the sense key is set to Hardware Error and the sense code is set to SCSI Parity Error.

Also in this embodiment, if a parity error continues to occur, the Ch transmitted at step 504
Since it is very unlikely that the eck Condition status will be received by the initiator, it is checked in step 505 whether the eck Condition status has been successfully transmitted (received by the initiator). If the message is not received normally, that is, the initiator has transmitted an Initiator Detected Error message or has transmitted some message, but has detected a parity error and has not been able to receive the message normally, a Command Complete message is transmitted. Instead, the process immediately shifts to the bus free phase in step 507, and the process ends.

As described above, in this embodiment, the SCSI
During a retry process started by the target detecting a parity error during the processing of the command, or during a retry process started by the initiator transmitting a report indicating that the parity error was detected to the target,
If a parity error is detected again or a detection report is received, the Check Condition status is transmitted to the initiator.If the Check Condition status cannot be transmitted normally, the bus free state is transmitted without transmitting a Command Complete message.・ In order to shift to the phase, even if a parity error that occurs continuously unless the cause of the error itself, such as a communication line error such as a data line disconnection, is eliminated, the error is immediately restored. Becomes possible. Further, it is possible to avoid the problem that the initiator and the target repeatedly exchange information indicating that a parity error has been detected between the initiator and the target and endlessly repeat retries that are unlikely to return.

[0038]

As described above, according to the present invention, even if a parity error such as an error in a communication line such as a data line disconnection occurs continuously unless the cause of the error itself is removed, the error can be recovered. This makes it possible to solve the problem that a retry that is unlikely to return as in the past is repeated indefinitely.

[Brief description of the drawings]

FIG. 1 is a first diagram based on the SCSI data transfer method of the present invention.
5 is a flowchart illustrating a flow of a process according to the embodiment.

FIG. 2 is a flowchart showing a processing flow of a second embodiment of the SCSI data transfer method of the present invention.

FIG. 3 is a flowchart showing a processing flow of a third embodiment of the SCSI data transfer system of the present invention.

FIG. 4 is a flowchart showing a processing flow of a fourth embodiment of the SCSI data transfer system of the present invention.

FIG. 5 is a flowchart showing a processing flow of a fifth embodiment of the SCSI data transfer system of the present invention.

FIG. 6 shows the structure of a logical block in a hard disk drive .
FIG .

FIG. 7 is a diagram showing a configuration of a general SCSI system;
There is .

──────────────────────────────────────────────────続 き Continuation of the front page (56) References JP-A-3-201051 (JP, A) JP-A-4-129450 (JP, A) JP-A-4-112254 (JP, A) JP-A-2- 291682 (JP, A) (58) Field surveyed (Int. Cl. ⁷ , DB name) G06F 13/00 301 G06F 13/10 310 G11B 20/10

Claims (1)

(57) [Claims] An information processing system for exchanging information between an initiator and a target based on the Small Computer System Interface (hereinafter abbreviated as SCSI) standard.
In the CSI data transfer method, during processing of a SCSI command, retry processing started by detecting a parity error by the target, or started by transmitting a report to the target that the initiator detected a parity error to the target. During the retry process, if a parity error is detected again or a report of the detection is received, the check condition status is transmitted to the initiator.If the check condition status cannot be transmitted normally, the command A SCSI data transfer method characterized by shifting to a bus free phase without transmitting a complete message.