FatFs - f_expand

f_expand

The f_expand function prepare or allocates a contiguous data area to the file.

FRESULT f_expand (
  FIL*    fp,  /* [IN] File object */
  FSIZE_t fsz, /* [IN] File size expanded to */
  BYTE    opt  /* [IN] Operation mode */
);

Parameters

fp: Pointer to the open file object.
fsz: Number of bytes in size to prepare or be allocated to the file. The data type FSIZE_t is an alias of either DWORD(32-bit) or QWORD(64-bit) depends on the configuration option _FS_EXFAT.
opt: Operation mode. Prepare only (0) or Allocate now (1).

Return Values

FR_OK, FR_DISK_ERR, FR_INT_ERR, FR_INVALID_OBJECT, FR_DENIED, FR_TIMEOUT

Description

The f_expand function prepare or allocates a contiguous data area to the file. When opt is 1, the function allocates a contiguous data area to the file. Unlike file expansion of file by f_lseek function, the file must be truncated prior to use this function and read/write pointer of the file stays at top of the file after the function. The file content allocated with this function is undefined. The function can be failed with FR_DENIED due to some reasons below.

No free contiguous space for the file was found.
Size of the file was not zero on expand.
The file has been opened in read-only mode.
Not allowable file size. (>= 4GiB on FAT volume)

When opt is 0, the function finds a contiguous data area and set it as suggested allocation point instead of allocating it to the file. The next cluster allocation is started at top of the contiguous data area. Thus the write file is guaranteed be contiguous and no allocation delay at least until the size reaches that size unless any other operation to the volume with changes of FAT is performed.

The contiguous file would have an advantage at time-critical read/write operations. It reduces some overheads in the file system and the storage media caused by random access due to fragmented file data. Especially at the exFAT volume, any FAT access for the contiguous file is completely eliminated and storage media will be accessed sequentially.

Also the contiguous file data can be accessed directory via low-level disk functions but it is not recommended in consideration for future compatibility.

QuickInfo

Available when _USE_EXPAND == 1 and _FS_READONLY == 0.

Example

    /* Creating a contiguous file */

    /* Create a new file */
    res = f_open(fp = malloc(sizeof (FIL)), "file.dat", FA_WRITE|FA_CREATE_ALWAYS);
    if (res) { /* Check if the file has been opened */
        free(fp);
        ...
    }

    /* Alloacte a 100 MiB of contiguous area to the file */
    res = f_expand(fp, 104857600, 1);
    if (res) { /* Check if the file has been expanded */
        free(fp);
        ...
    }
    /* Now you have a contiguous file accessible with fp */

    /* Accessing the file data directly via low-level disk functions */

    /* Get physical location of the file data */
    drv = fp->obj.fs->drv;
    sect = fp->obj.fs->database + fp->obj.fs->csize * (fp->obj.sclust - 2);

    /* Write 2048 sectors from top of the file at a time */
    res = disk_write(drv, buffer, sect, 2048);