JPH01196623A - Self-diagnosing method for disk controller - Google Patents

Abstract

PURPOSE:To execute the self-diagnosis of a memory in a sector buffer without giving a bad influence to a data input and output processing by judging the propriety of a data buffer, for example, the sector buffer at the idle time of a disk controller. CONSTITUTION:Supervision processing function 6 of a request command and memory diagnosis processing function 10 are equipped in a disk controller 4. The request supervision processing function 6 supervises the request command from a computer side and executes the processing excepting for the original function of the disk controller. For example, the function 6 judges whether a memory diagnosis processing can be activated or not. When the request supervision processing function 6 judges that a processing request is not generated from the computer side for a while, the memory diagnosis processing function 10 is activated and the propriety judgement of memories 4A and 4B in the sector buffer is executed. Thus, the self-diagnosis of the sector buffer memory in the disk controller can be executed.

Description

DETAILED DESCRIPTION OF THE INVENTION <Field of Industrial Application> The present invention relates to a method for self-diagnosing the presence or absence of an abnormality in a data buffer in a disk controller without adversely affecting high-speed processing of the disk itself.

<Conventional technology> The disk controller is usually a CPU (Central Processing Unit).
It stands between the disk body (disk drive device) and sends and receives data between the two, and since the storage disk of the disk body rotates at extremely high speed, a sector buffer is used as a data buffer to keep up with this speed. We have multiple units and perform high-speed data input/output processing.

Furthermore, it is required to perform input/output data processing with high reliability.

Normally, a disk input/output request program on the CPU side sets four sectors of output data on a DMA (direct memory access) buffer with the disk controller, then sets a request command on the command buffer, and then sends a request command to the disk controller. Generates an interrupt to the controller.

The disk collector uses an interrupt to check the contents of the command buffer, operates according to the type of command, and operates the disk itself. In addition, all input/output data is relayed by a sector buffer, and DMA with the CPU is
In parallel with the transfer, input/output is performed with sectors on the storage disk of the disk itself.

As mentioned above, the disk controller is designed with the main focus on data input/output to the storage disk, and since it only has multiple sector buffers to perform data input/output processing, it is capable of performing other functions than data input/output processing. It is difficult to incorporate processing functions and usually does not have diagnostic functions.

That is, with the exception of disk devices connected to large combinators, ordinary disk controllers do not have parity error checks or E
It does not have diagnostic functions such as CC (error correction).

However, adding a diagnostic function would result in overhead, and the data input/output processing would not be able to keep up with the high-speed rotation of the storage disk, and since the sector buffer is also in use, there is a limit to how much data can be checked.

<Problem to be solved by the invention> However, data reliability is important for disk devices, and if the sector buffer that relays all input/output data becomes defective, it may notify the CPU of incorrect data. There is a risk of writing to the disc itself.

SUMMARY OF THE INVENTION In view of the above-mentioned prior art, it is an object of the present invention to provide a method for self-diagnosing the quality of a data buffer of a disk controller without adversely affecting data input/output processing.

Means for Solving the Problems> A self-diagnosis method for a disk controller according to the present invention includes the following steps:
When data input/output commands from the computer side are interrupted for a certain period of time, a specific pattern of data is written to the data buffer and read out for comparison before the next data input/output command occurs, and if the data does not match, the data is It is characterized by determining that the buffer is defective.

When we investigated the frequency of occurrence of data input/output commands from the computer side, we found that normally, once data input/output commands occur, they occur continuously, but conversely, once they are interrupted, it takes a long time for the next one to occur. , it turns out that there is enough headroom to perform memory diagnostics on data buffers, e.g. sector buffers.

Therefore, we monitor request commands from the computer side, and if data input/output commands are interrupted for a certain period of time,
Since there will be no processing requests from the computer for the time being, it can be determined that there is no problem in starting processing other than its original function. Therefore, if data input/output commands are interrupted for a certain period of time, by writing a specific pattern of data to the data buffer, reading it, and comparing it, it is possible to self-diagnose the quality of the data buffer without adversely affecting data input/output processing. can.

Note that the self-diagnosis function of the data buffer may be activated at regular intervals, or may be interrupted if a request command is generated from the computer side during the self-diagnosis process.

Embodiment One embodiment of the present invention will be described with reference to FIGS. 1 and 2.

Figure 1 shows the flow of data inside the disk controller,
The relationship between the computer side and sectors of the storage disk is shown, and FIG. 2 shows a concrete implementation method of the present invention.

In FIG. 1, as mentioned above, when the disk input/output request program 1 on the computer side normally sets output data for four sectors A to D on the memory 2 of the DMA buffer with the disk controller 4, A request command is set in the memory 3 of the command buffer, and an interrupt is generated in the disk controller 4. In disk controller 4, command buffer 3 is accessed by interrupt.
Check the data type and use functions such as data input/output processing 5 depending on the type. The disk controller 4 has a plurality of sector buffers 4A and 4B to keep up with the high-speed rotation of the storage disk 11 of the disk main body, and all input and output data is intermediated in the sector buffers 4A and 4B and communicated with the computer side. A on the memory disk 11 in parallel with DMA transfer
- Performs data input/output with the four sectors 12 of D. In the figure, ■ to ■ indicate the flow of data.

Furthermore, the disk controller 4 has a request command monitoring processing function 6 and a memory diagnosis processing function 10. The request monitoring processing function 6 monitors request commands from the computer side and determines whether or not it is appropriate to start processing other than the original function of the disk controller, here memory diagnostic processing. When the request monitoring processing function 6 determines that no processing request will be issued from the computer side for the time being, it starts the memory diagnostic processing function 10 and makes it judge whether the memos 1J4A and 4B in the sector buffer are good or bad.

Specifically, as shown in Figure 2, the request monitoring processing function is as follows:
It consists of a command analysis process 7 that corresponds to an interrupt from the computer side, and an interval timer process 8 that is activated by a timer interrupt at a fixed period.

In the command analysis process 7, the command is activated by a request from the computer side, analyzes whether the request is a normal input/output command, and if it is an input/output command, the data input/output processing function 5 is activated as usual. . In addition, if it is an input/output command, interval timer processing 8 is executed for each request.
Clear the monitoring timer 9 to zero.

In the interval timer processing 8, the monitoring timer 9 is counted up each time it is started, and the value is set to a constant value 9.
When the above is indicated, it is determined that the processing request from the Combi Coater side has ceased, and the memory diagnostic processing function 10 is activated for the first time. After that, if there is no further processing request, that is, if the monitoring timer 9 is not cleared, the memory diagnostic processing function 10 continues to be activated. In addition, if a processing request is generated from the computer side during the process and the monitoring timer 9 is cleared,
The memory diagnostic processing function 10 is to be canceled. At this time, the data input/output processing function 5 is activated and responds to the request from the computer side.

When activated, the memory diagnostic processing function 10 writes a specific pattern of data to the sector buffer 4A and/or 4B, reads it out again, and compares the data to determine whether the sector buffer is good or bad. Needless to say, if the written data and the read data do not match, it is determined to be defective. If it is determined to be defective, the status is retained and the next time a processing request is issued from the computer side, a response to the computer side indicating an abnormality is made.

On the computer side, by receiving an abnormal response,
Processing can be stopped immediately to prevent data from becoming fraudulent.

Note that after the memory diagnostic process is stopped, the next time the interval timer process 8 starts, the diagnosis can be restarted from the stopped address, or all addresses in the sector buffer can be divided into multiple blocks and diagnosed block by block. , the diagnosis may be restarted from the block where it was stopped.

In the above embodiment, the interval timer process 8 is activated by a timer interrupt at a fixed period, and the monitoring timer 9 is counted up each time it is activated, but the monitoring timer 9 may be constantly counted up.

Furthermore, the processes 6, 7, 8, 9 and 10 described above may be realized by either software or hardware, or may be used in combination.

<Effects of the Invention> According to the present invention, since the quality of the data buffer, such as the sector buffer, is determined when the disk controller is completely idle (idle), no adverse effects such as slowing down of data input/output processing occur.

In addition, since the memory of the sector buffer, which is used most frequently in the disk controller, can be diagnosed almost constantly,
The reliability of input and output data is greatly improved.

Therefore, the present invention is an extremely effective diagnostic method for disk controllers that require high-speed processing and high data reliability.

[Brief explanation of the drawing]

Figure 1 shows the flow of data inside the disk controller,
FIG. 2 is a diagram showing the relationship between the computer side and sectors of the storage disk, and FIG. 2 is a diagram showing a specific example of implementing the present invention. In the drawing, 1 is a disk input/output request program, 2 is a DM
A buffer, 3 is a command buffer, 4 is a disk controller, 4A and 481 are sector buffers, 5
is a data input/output processing function, 6 is a request monitoring processing function, 7 is a command analysis process, 8 is an interval timer process, 9
1 is a monitoring timer, and 10 is a memory diagnostic processing function.

Claims (1)

[Claims] When data input/output commands from the computer side are interrupted for a certain period of time, a specific pattern of data is written to the data buffer and read out for comparison before the next data input/output command occurs, and if the data does not match, the data is A self-diagnosis method for a disk controller characterized by determining that a buffer is defective.