BDOS File I/O Services

Thursday, 26 March 2009 10:19 Shael
Print

The BDOS File I/O Services allow a user application the ability to transfer data blocks to and from a CPM file. CPM files can be accessed either in sequential mode or in random access mode. The BDOS File I/O Services require that the file be opened or created before data may be transferred. See the article on the BDOS file services for more information about accessing and creating files. Many of the service functions relating to file I/O service require the memory address of an FCB (File Control Block) structure. The following information was extracted from the Digital Research CPM2.2 manual.

Function 20: Read Sequential
RegisterValue
Entry C 14H
DE FCB Address
Return A Directory Code

Given that the FCB addressed by DE has been activated through an Open oe Make file function, the Read Sequential function reads the next 128-byte record from the file into memory at the current DMA address. The record is read from position cr of the extent, and the cr field is automatically incremented to the next record position. If the cr field overflows, the next logical extent is automatically opened and the cr field is reset to zero in preparation for the next read operation. The value 00H is returned in the A register if the read operation was successful, while a nonzero value is returned if no data exist at the next record position (for example, end-of-file occurs).

Function 21: Write Sequential
RegisterValue
Entry C 15H
DE FCB Address
Return A Directory Code

Given that the FCB addressed by DE has been activated through an Open or make function, the Write Sequential function writes the 128-byte data record at the current DMA address to the file named by the FCB. The record is placed at position cr of the file, and the cr field is automatically incremented to the next record position. If the cr field overflows, the next logical extent is automatically opened and the cr field is reset to zero in preparation for the next write operation. Write operations can take place into an existing file, in which case newly written records overlay those that already exist in the file. Register A = 00H upon return from a successful write operation, while a nonzero value indicates an unsuccessful write caused by a full disk.

Function 26: Set DMA Address
RegisterValue
Entry C 1AH
DE DMA Address
Return (none)

DMA is an acronym for Direct Memory Address, which is often used in connection with disk controllers that directly access the memory of the mainframe computer to transfer data to and from the disk subsystem. Although many computer systems use non-DMA access (that is, the data is transferred through programmed I/O operations), the DMA address has, in CP/M, come to mean the address at which the 128-byte data record resides before a disk write and after a disk read. Upon cold start, warm start, or disk system reset, the DMA address is automatically set to BOOT+0080H. The Set DMA function can be used to change this default value to address another area of memory where the data records reside. Thus, the DMA address becomes the value specified by DE until it is changed by a subsequent Set DMA function, cold start, warm start, or disk system reset.

Function 34: Write Random
RegisterValue
Entry C 22H
DE FCB Address
Return A Return Code

The Write Random operation is initiated similarly to the Read Random call, except that data is written to the disk from the current DMA address. Further, if the disk extent or data block that is the target of the write has not yet been allocated, the allocation is performed before the write operation continues. As in the Read Random operation, the random record number is not changed as a result of the write. The logical extent number and current record positions of the FCB are set to correspond to the random record that is being written. Again, sequential read or write operations can begin following a random write, with the notation that the currently addressed record is either read or rewritten again as the sequential operation begins. You can also simply advance the random record position following each write to get the effect of a sequential write operation. Note that reading or writing the last record of an extent in random mode does not cause an automatic extent switch as it does in sequential mode.

The error codes returned by a random write are identical to the random read operation with the addition of error code 05, which indicates that a new extent cannot be created as a result of directory overflow.

Function 35: Compute File Size
RegisterValue
Entry C 23H
DE FCB Address
Return Random Record Field Set

When computing the size of a file, the DE register pair addresses an FCB in random mode format (bytes r0, r1, and r2 are present). The FCB contains an unambiguous filename that is used in the directory scan. Upon return, the random record bytes contain the virtual file size, which is, in effect, the record address of the record following the end of the file. Following a call to Function 35, if the high record byte r2 is 01, the file contains the maximum record count 65536. Otherwise, bytes r0 and r1 constitute a 16-bit value as before (r0 is the least significant byte), which is the file size.

Data can be appended to the end of an existing file by simply calling Function 35 to set the random record position to the end-of-file and then performing a sequence of random writes starting at the preset record address.

The virtual size of a file corresponds to the physical size when the file is written sequentially. If the file was created in random mode and holes exist in the allocation, the file might contain fewer records than the size indicates. For example, if only the last record of an 8-megabyte file is written in random mode (that is, record number 65535), the virtual size is 65536 records, although only one block of data is actually allocated.

Function 36: Set Random Record
RegisterValue
Entry C 24H
DE FCB Address
Return Random Record Field Set

The Set Random Record function causes the BDOS automatically to produce the random record position from a file that has been read or written sequentially to a particular point. The function can be useful in two ways.

First, it is often necessary initially to read and scan a sequential file to extract the positions of various key fields. As each key is encountered, Function 36 is called to compute the random record position for the data corresponding to this key. If the data unit size is 128 bytes, the resulting record position is placed into a table with the key for later retrieval. After scanning the entire file and tabulating the keys and their record numbers, the user can move instantly to a particular keyed record by performing a random read, using the corresponding random record number that was saved earlier. The scheme is easily generalized for variable record lengths, because the program need only store the buffer-relative byte position along with the key and record number to find the exact starting position of the keyed data at a later time.

A second use of Function 36 occurs when switching from a sequential read or write over to random read or write. A file is sequentially accessed to a particular point in the file, Function 36 is called, which sets the record number, and subsequent random read and write operations continue from the selected point in the file.

Function 40: Write Random with Zero Fill
RegisterValue
Entry C 28H
DE FCB Address
Return A Return Code

The Write With Zero Fill operation is similar to Function 34, with the exception that a previously unallocated block is filled with zeros before the data is written.

Last Updated on Friday, 06 November 2009 20:23