JPH0158580B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention provides read after write operations.
This invention relates to a magnetic tape control method in a device that performs writing.

Conventionally, a magnetic tape device has a read head after a write head, and when writing data, the data is read immediately after writing to check whether the data has been correctly written to the magnetic tape. When an error is detected in this check, the magnetic tape is rewound to the beginning of the error block and the data is rewritten.

However, with this method, the magnetic tape must be rewound and data must be rewritten every time an error is detected, which not only increases the write processing time but also
There have been various problems such as destruction of a normal data block before an error block or failure to erase an error block that should be erased due to the rewinding operation. Furthermore, even if an error mark is written to a data block in which an error was detected during a read check during a write operation, there is a risk that this error mark will be misjudged during a read operation. An increase will occur.

In view of these points, the present invention reduces the data write processing time by not rewinding the magnetic tape at all during the write operation, and also by writing not only the error mark but also the block number in the data block. The present invention provides a new magnetic tape control method that allows error blocks to be reliably determined during a read operation.

Hereinafter, the present invention will be explained with reference to the drawings.

FIG. 1 is a block diagram showing an embodiment of the magnetic tape control system according to the present invention, in which 1 is a magnetic tape on which a data block D is recorded, 2 and 3 are a read head and a write head, respectively, and 4 is a magnetic tape. 5 is a processing unit that performs various processes; 6 is a control signal RSC based on data from the processing unit 5;
7 is an input/output control unit that outputs FBC and inputs control signals TMO and BLS from the magnetic tape device 4; 7 is a serial-parallel conversion unit that mutually converts parallel data to serial data; and serial data to parallel data; 8; 9 is a modulation/demodulation unit that modulates write data and demodulates read data; 9 is an address decoder that selects either the input/output control unit 6 or the serial/parallel conversion unit 7 according to the address from the processing unit 5;
1 and 12 are a block number buffer, a write data buffer, and a read data buffer, respectively.

Next, the operation of this embodiment will be explained.

First, in a write operation, the input/output control section 6 is selected by applying a write address from the processing section 5 to the address decoder 9, data from the processing section 5 is taken into the input/output control section 6, and the tape is run. A signal RSC instructing the tape running direction and a tape running direction switching signal FBC are output. These signals drive the magnetic tape device 4, so that the magnetic tape 1 is run in the forward direction.

Subsequently, the processing section 5 gives an address to the address decoder 9 to select the serial/parallel converter 7, takes in data from the block number buffer 10 and the write data buffer 11, and sets it in the serial/parallel converter 7 via the data bus. do. The serial/parallel converter 7 converts the input parallel data into serial data, and this data is modulated by the modulator/demodulator 8 and applied to the magnetic tape device 4 as write data WDT. The block number and data are written in as shown in FIG. 2D.

Here, the block number buffer 10 is the second
As shown in Figure A, it constitutes a so-called register, and the block numbers are 1, 2, 3, . . . and increase in order, corresponding to the data blocks. The write data buffer 11 is a buffer for storing one block of data, as shown in FIG. In this embodiment, in order to read and check data, the lower two bytes of the summed data from data 1 to data n are written after data n as checksum data, as shown in FIG. 2D.

By the way, when data is written, the written data is read by the reading head 2 using the read-after-write method and is applied from the magnetic tape device 4 to the modulation/demodulation section 8 as read data RDT. The demodulated read data is then converted from serial data to parallel data in a serial/parallel converter 7. At this time, the serial/parallel converter 7
Since the processing section 5 gives the interrupt signal IRQ to the address decoder 9, the address to select the serial/parallel conversion section 7 is sent to the address decoder 9. Therefore, the read data is sent to the read data buffer 12 via the processing section 5, as shown in the diagram (H) of FIG. It is stored as follows.

Therefore, the processing unit 5 uses the read data buffer 12.
data 1 to data n are added, and the lower two bytes of the added data are compared with the read checksum data. As a result, when the data match, GAP is written after the data block as shown in Figure 3A, and
Add 1 to the block number buffer 10,
The next write data is taken into the write data buffer 11 from the host 13, and the next data block is written onto the magnetic tape 1 by the same operation as described above.

However, if the checksum data does not match, it is determined that the data could not be written correctly for some reason, and an error mark EM, EM, etc. is placed after the data, as shown in data block B in Figure 3A. For example, write a few bytes of "FF",
After making the data block B an error block and creating a gap without updating the value of the block number buffer 10, write the same block number and the same data that should be written to the error block B again in the next data block B'. Write. however,
When writing the error mark EM, do not perform a read-after-write check. After writing data block B', as described above, the gap GAP is written, the block number is updated, the next write data is taken from the host 13 into the write data buffer 11, and the next block number is written to data block C. and write data.

In this manner, during writing, data is written one after another without rewinding the magnetic tape in order to rewrite the data block.

Next, the reading operation will be explained.

For read operations as well as for write operations,
First, the input/output control section 6 is selected by the processing section 5, and the control signals RSC and FBC are sent to the magnetic tape device 4.
In addition to this, the magnetic tape 1 is controlled to run in the forward direction. Next, the serial-to-parallel converter 7 is selected, and the data read by the reading head 2 is input to the modulation/demodulator 8 via the magnetic tape device 4, where it is demodulated and then sent to the serial-to-parallel converter 7.
is converted to parallel data. and interrupt signal
The read data is taken into the read data buffer 12 by IRQ.

Therefore, similarly to the read-after-write check, the lower two bytes of the data obtained by adding data 1 to data n that have been read are compared with the checksum data. As a result of comparison, if the data match,
Add 1 to the block number and read the next data block.

However, if the data do not match, an error mark is checked. A block signal BLS is used to confirm this error mark. That is, the block signal BLS is "L" during the data block.
The error mark EM is written at the end of the data block, that is, immediately before the gap GAP. Therefore, the block signal BLS is "L".
Error marks can be confirmed by checking the data that is about to change from the level to the "H" level.

As a result of the confirmation, if there is no error mark, this data block is read again and the check sum is again checked. That is, by sending control data from the processing section 5 to the input/output control section 6, the signal RSC is outputted and the signal FBC is inverted, so that the magnetic tape 1 is run in the back direction from the stop position P shown in FIG. 3B. let After the block signal changes from the "L" level to the "H" level, a constant count is performed using the signal TMO from the tachometer, and the magnetic tape is stopped at the Q position as shown in FIG. 3B. Here, the reason why the signal TMO is used is to stop the magnetic tape at a predetermined position in the gap GAP. When the magnetic tape stops, the signal FBC is inverted again to run the magnetic tape in the forward direction and read it again. If the checksum does not match even after re-reading several or tens of times, it is treated as a reading error.

However, as a result of checking the error mark, if there is an error mark, the next data block is read without rereading as described above. In other words, the presence of the error mark determines that this data block has not been written correctly.
Instead of rereading, the next data block containing normal data is read.

By the way, since error marks are also read in the same way as normal data, there is a possibility that error marks will be misjudged. That is, in FIG. 3A, it may be determined that a data block A without an error mark has an error mark, or it may be determined that a data block B with an error mark does not have an error mark.

However, in the present invention, sequentially increasing block numbers are written in each data block, and the same block number is written in the error block and the next data block. Even if it is determined that data block A has an error mark, the block number of the next data block B is greater than the block number of data block A.
It turned out to be a misjudgment of an error mark.
Furthermore, even if it is determined that there is no error mark in data block B, the block number of the next data block B' is the same as the block number of data block B, so it is still determined that the error mark is incorrectly determined. Ru. In this way, by using block numbers, the problem of erroneous determination of error marks is solved.

As described above, the magnetic tape control method according to the present invention does not rewind the magnetic tape during writing, so it can shorten the write processing time and prevents destruction of previous data and error blocks caused by rewinding. Various problems such as complete erasure can be solved.

Furthermore, during reading, it can be determined with certainty that a data block has an error based on the error mark and block number, which is effective in shortening the reading time.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing an embodiment of the magnetic tape control system according to the present invention, FIGS. 2A to 2C are explanatory diagrams showing the contents of various data buffers, and FIG. FIG. 3A is an explanatory diagram showing the state of writing a data block onto the magnetic tape, and FIG. 3B is a waveform diagram showing the block signal BLS. Explanation of main drawing numbers: 1...magnetic tape, 2...read head, 3...write head, 4...magnetic tape device, 5...processing section, 6...input/output control section, 7...direct Parallel conversion section, 8... Modulation/demodulation section, 9...
...Address decoder, 10...Block number buffer, 11...Write data buffer, 1
2...Read data buffer.

Claims (1)

[Claims] 1. In a device that adds a different block number to each data block, writes the data block on a magnetic tape across a gap, and reads the data block immediately after writing, it is possible to detect an error during a read check during the write operation. When detected, an error mark is written in the data block, and the same data and the same block number as the data to be written in the data block are written in the next data block. During a read operation, the presence or absence of the error mark is determined and . A magnetic tape control system characterized in that it is determined whether a block number in a read data block is different from a block number in a data block immediately before the read data block.