JPH0731640B2 - Disk access control method - Google Patents

Description

DETAILED DESCRIPTION [Overview] Regarding a disk access control method for writing data to a disk device, even if a power failure occurs while writing data to the disk device, the data saved when the power is turned on again is restored. The data to be written to the disk device is stored for the purpose of improving reliability by rewriting the ECC or CRC while writing, and at the same time, the unwritten flag indicating whether or not this data was written to the disk device is stored. The non-write flag is turned on after the data transferred from the write request source is stored in the non-volatile memory and the power failure occurs while writing this data to the disk device. At the time of re-input, the data whose unwritten flag is on is written to the disk device.

[Industrial application field]

The present invention relates to a disk access control method for writing data in a disk device.

[Problems to be solved by conventional technology and invention]

Conventionally, in the case of the computer system as shown in FIG. 4A, the disk device 13 is accessed by an instruction from the CPU 15. For example, if a power failure such as a momentary interruption occurs during writing, one sector cannot be completely written, as shown in Fig. 4 (b), and ECC, CRC, etc., that writes a part of the data and the end of the sector. A state occurs in which the check bit cannot be written. Then, when the power is turned on again, if part of this data and data lacking check bits such as ECC and CRC are read, an error occurs and a problem such as a system down occurs. was there. The following is a brief description of FIG.

In FIG. 4A, the data transferred via the reselection connection unit 17 is stored in the MEM (memory) 14. This MEM
The data stored in 14 is written in the disk device 13 under the control of the DISK controller (disk controller) 11. Then, if necessary, the data read from the disk device 13 is written in the MEM 14 and then the line connection unit 1
It is sent to the terminal etc. via 7.

In this case, as shown in FIG. 4B, if a power failure such as a momentary interruption of the AC input occurs while data is being written to the disk device 13, the ECC of part or the end of the data will be lost. There is a situation where check bits such as CRC are not written. Therefore, when data is read from this sector when the power is turned on again, there is a problem that an error is detected by check bits such as ECC and CRC, and a system down occurs.

According to the present invention, even if a power failure occurs while writing data to the disk device, the data saved when the power is turned on again is rewritten and the ECC or CRC is regenerated to improve reliability. Has an aim.

[Means for solving problems]

 FIG. 1 shows the principle configuration of the present invention.

In FIG. 1, the disk controller 1 is a disk device 3
Access control.

The non-volatile memory 2 holds stored data and the like even when the power is cut off, and stores the control table 2-1 and the data 2-2.

The control table 2-1 includes a disk address (address to be stored in the disk device 3) of the data 2-2 stored in the volatile memory 2, an unwritten flag (a flag indicating whether or not the data has been written to the disk device 3). ), An NVM address (an address in the nonvolatile memory 2 in which data is stored), and the like are stored.

Data 2-2 is data to be written to the disk device 3.

The disk device 3 stores data.

The MEM (memory) 4 stores data transferred from a terminal or the like and stores data read from the disk device 3.

The CPU 5 performs various controls.

[Action]

As shown in FIG. 1, the present invention stores the data read from the MEM4 as the data 2-2 in the non-volatile memory 2 and then sets the unwritten flag to the ON state to
A power failure occurred while writing 2 to the disk device 3,
When the power is turned on again, the data 2-2 corresponding to the unwritten flag turned on is written to the disk device 3.

Therefore, even if a power failure such as a momentary interruption occurs while writing the data 2-2 to the disk device 3, it becomes possible to write the saved data to the disk device 3 when the power is turned on again. The reliability of can be improved.

〔Example〕

Next, the configuration and operation of one embodiment of the present invention will be sequentially described in detail with reference to FIG. 2 and FIG.

In FIG. 2, the disk control unit 1 includes 1-1 to 1
-4, 2-3, 2-4, and controls access to the disk device 3.

The P-mode DMA interrupt control unit 1-1 uses the P-mo of the common bus.
It controls de, DMA and interrupts.

The disk control unit 1-2 controls the disk device 3 to perform access control.

ROM (read-on memory) 1-3 is a disk controller 1
It stores various programs for controlling.

The MPU1-4 performs various controls according to the program read from the ROM1-3.

The RAM (random access memory) 2-3 and the NVM (nonvolatile memory) 2-4 are backed up by a backup battery 6, and the control table 2-1 and the data stored in them are stored even when the power is cut off. 2-2 is held. It should be noted that instead of using the non-volatile memory 2 by backing up the power supply by the backup battery 6 such as the RAM 2-3, use any readable / writable element having a function of saving data even when the power is cut off. Good.

The disk device 3 stores data.

The MEM (memory) 4 stores data and the like.

The CPU 5 performs various controls.

The backup battery 6 backs up the power sources of the RAM 2-3 and NVM.

The line connection unit 7 connects to the line.

Next, the operation of the configuration shown in FIG. 2 will be described in detail with reference to FIG.

In FIG. 3, the figure shows a state in which data is transferred from the memory to the NVM by DMA. This means that the P-mode DMA interrupt control unit 1-1 is activated by the notification from the CPU 5, and the data read from the MEM 4 is DMA-transferred to the NVM 2-4 and written (FIG. 1 data 2). -2). At this time,
In the RAM 2-3, as shown in the control table 2-1 in FIG. 1, the disk address for writing the data to the disk device 3 and the NVM in the NVM 2-4 in which the data is stored.
An address and an unwritten flag indicating whether or not the data is written in the disk device 3 are stored.

The figure shows a state in which various addresses and unwritten flags are set in the RAM 2-3.

In the figure, the state of writing to the disk is shown. this is,
RAM2 whose power is backed up by backup battery 6
From the control table 2-1 stored in -3
The NVM address is read, and the data read from the NVM2-4 by this read NVM address is transferred to the disk address in the disk device read from the control table 2-1 stored in the RAM2-3. This means writing with check bits such as ECC and CRC added to each sector.

The figure shows a state in which the data whose unwritten flag is on is written to the disk device when the power is turned on again. this is,
In the figure, the control table 2-corresponding to the occurrence of a power failure while writing data to the disk device 3
Since the unwritten flag in 1 is saved in the ON state, the data saved in NVM2-4 corresponding to the ON state of this unwritten flag is stored in the disk device 3 after the power is turned on again. Means to write to. After writing the data to the disk device 3, the unwritten flag is turned off.

As described above, after storing the data 2-2 and the control table 2-1 in the NVM2-4 and the RAM2-3 which are backed up by the backup battery 6 in the disk control unit 1, the unwrite flag is turned on, Data 2-
2 is written to the disk device 3, the power is cut off, and when the power is turned on again, the data corresponding to the unwritten flag turned on is written to the disk device 3 to prevent the data from being destroyed and improve the system reliability. It becomes possible.

〔The invention's effect〕

As described above, according to the present invention, the non-volatile memory is provided in the disk control unit, the data to be written to the disk device and the control information are stored in the non-volatile memory, and then the unwritten flag is turned on to save the data. If a power failure occurs during writing to the disk device, a write error may occur.Therefore, the data is written to the disk device when the power is turned on again. Even if a power failure such as occurs, save the data being written to the disk device,
In addition, it is possible to improve the reliability of the system by avoiding a system down caused by the occurrence of a read error by rewriting when the power is turned on again.

[Brief description of drawings]

FIG. 1 is a block diagram showing the principle of the present invention, FIG. 2 is a block diagram showing an embodiment of the present invention, and FIG. 3 is a flowchart for explaining the operation of the present invention.
FIG. 4 shows an explanatory view of the prior art. In the figure, 1 is a disk controller, 2 is a non-volatile memory, 2-
Reference numeral 1 is a control table, 2-2 is data, 3 is a disk device, 4 is MEM (memory), and 5 is a CPU.

Claims (1)

[Claims] 1. In a disk access control method for writing data to a disk device, whether the data to be written is stored in the disk device (3) and the data is written to the disk device (3) in units of one sector or a plurality of sectors. A non-volatile memory (2) for storing a non-write flag indicating whether or not the data is transferred from the write request source is stored only in the non-volatile memory (2) while the power is on. After turning on, writing a check bit for each sector to each sector to write one or more sectors to the disk device (3) and then turning off the unwritten flag in units of one sector or multiple sectors, When a power failure occurs, this unwritten flag remains on when the power is turned on again. Disk access control method which is characterized by being configured to the data of a sector or a plurality sectors in the non-write flag OFF state after writing to the disk device (3).