The FatFs module assumes that size of <tt>char</tt>/<tt>short</tt>/<tt>long</tt> are 8/16/32 bit and <tt>int</tt> is 16 or 32 bit. These correspondence are defined in <tt>integer.h</tt>. This will not be a problem on most compilers. When a conflict with existing definitions is occured, you must resolve it with care.</li>
<p>(a) If a working disk module with FatFs disk interface is provided, nothing else will be needed. (b) To attach existing disk drivers with different interface, glue functions are needed to translate the interfaces between FatFs and the drivers.</p>
<p>You need to provide only low level disk I/O functions required by FatFs module and nothing else. If a working disk module for the target system is already existing, you need to write only glue functions to attach it to the FatFs module. If not, you need to port another disk module or write it from scratch. Most of defined functions are not that always required. For example, disk write function is not required at read-only configuration. Following table shows which function is required depends on the configuration options.</p>
<tr><td>disk_status<br>disk_initialize<br>disk_read</td><td>Always</td><tdrowspan="5">Disk I/O functions.<br>Samples available in ffsample.zip.<br>There are many implementations on the web.</td></tr>
<tr><td>ff_cre_syncobj<br>ff_del_syncobj<br>ff_req_grant<br>ff_rel_grant</td><td>_FS_REENTRANT == 1</td><tdrowspan="2">O/S dependent functions.<br>Samples available in option/syscall.c.</td></tr>
<p>These are the memory usage on some target systems with following condition. The memory sizes are in unit of byte, <em>V</em> denotes option <tt>_VOLUMES</tt> and <em>F</em> denotes number of open files. All samples here are optimezed in code size.</p>
<p>Follwing table shows which API function is removed by configuration options for the module size reduction. To use any API function, the row of the function must be clear.</p>
<p>FatFs module supports long file name (LFN). The two different file names, short file name (SFN) and LFN, of a file is transparent on the API except for <tt>f_readdir</tt> function. The support for LFN is disabled by default. To enable the LFN, set <tt><ahref="config.html#use_lfn">_USE_LFN</a></tt> to 1, 2 or 3, and add <tt>option/unicode.c</tt> to the project. The LFN requiers a certain working buffer in addition. The buffer size can be configured by <tt><ahref="config.html#max_lfn">_MAX_LFN</a></tt> according to the available memory. The length of an LFN will be up to 255 characters, so that the <tt>_MAX_LFN</tt> should be set to 255 for all file names. If the size of working buffer is insufficient for the input file name, the file function fails with <tt>FR_INVALID_NAME</tt>. When use any re-entry to the API with LFN is enabled, <tt>_USE_LFN</tt> must be set to 2 or 3. In this case, the file function allocates the working buffer on the stack or heap. The working buffer occupies <tt>(_MAX_LFN + 1) * 2</tt> bytes and additional <tt>(_MAX_LFN + 44) / 15 * 32</tt> bytes when exFAT is enabled.</p>
<p>When the LFN is enabled, the module size will be increased depends on the configured code page. Right table shows how much code size increased when LFN is enabled with some code pages. Especially, in the CJK region, tens of thousands of characters are being used. Unfortunately, it requires a huge OEM-Unicode bidirectional conversion table and the module size will be drastically increased as shown in the table. As the result, the FatFs with LFN enebled with those code pages will not able to be ported on the most 8-bit MCU systems.</p>
<p>Note that the support for LFN on the FAT volume is a patent of Microsoft Corporation. This is not the case on FAT32 but most FAT32 drivers come with the support for LFN. FatFs can switch the LFN on/off by configuration option. When enable LFN on the commercial products, you will need to be licensed by Microsoft depends on the final destination of the products.</p>
<p>By default, FatFs uses ANSI/OEM code set on the API even at LFN configuration. FatFs can also switch the character encoding on the API to Unicode by configuration option <tt><ahref="config.html#lfn_unicode">_LFN_UNICODE</a></tt>. This means that FatFs supports the full featured LFN specification. The data type <tt>TCHAR</tt> specifies strings on the API is an alias of <tt>char</tt>(ANSI/OEM) or <tt>WCHAR</tt>(UTF-16). For more information, refer to the description in the <ahref="filename.html#uni">file name</a>.</p>
<p>The exFAT (Microsoft's Extended File Allocation Table) file system is a replacement of the FAT file system which has been widely used in the embedded systems and consumer devices. It is adopted by SDA (SD Association) as a recommended file system for high capacity SD cards (>32GB) and they are being shipped with this format, so that the exFAT become one of the standard file systems for removable media as well as FAT. The exFAT file system allows the file size larger than 4 GiB limit what FAT file system allows upto and some file system overhead, especially cluster allocation delay, are reduced as well. This feature improves the write throughput to the file.</p>
<p>Note that the exFAT is a patent of Microsoft Corporation. The exFAT function of FatFs is an implementation based on <cite>US. Pat. App. Pub. No. 2009/0164440 A1</cite>. FatFs module can swich the exFAT on/off by configuration option. When enable the exFAT on the commercial products, you will need to be licensed by Microsoft depends on the final destination of the products.</p>
<p><em>Remark: Enabling exFAT discards ANSI C (C89) compatibility because of need for 64-bit integer type.</em></p>
<p>The file operations to the <em>different volume</em> each other is always re-entrant regardless of configurations except when LFN enabled with static working buffer. It can work simultaneously without any mutual exclusion.</p>
<p>The file operations to the <em>same volume</em> is not re-entrant. It can also be configured thread-safe by option <tt><ahref="config.html#fs_reentrant">_FS_REENTRANT</a></tt>. It enables to control exclusive use of each file system object. In this case, also the OS dependent synchronization object control functions, <tt>ff_cre_syncobj/ff_del_syncobj/ff_req_grant/ff_rel_grant</tt>, needed to be added to the project. There are some examples in the <tt>option/syscall.c</tt>.</p>
<p>When a file function is called while the volume is being accessed by other task, the file function to the volume will be suspended until that task leaves the file function. If the wait time exceeded a period defined by <tt>_TIMEOUT</tt>, the file function will abort with <tt>FR_TIMEOUT</tt>. The timeout function might not be supported on the some RTOSs.</p>
<p>There is an exception on the re-entrancy for <tt>f_mount/f_mkfs/f_fdisk</tt> function. These volume management functions are not re-entrant on the same volume and corresponding physical drive. When use these functions, other tasks need to avoid to access the volume.</p>
<p>Note that this section describes on the re-entrancy of the FatFs module itself. The <tt>_FS_REENTRANT</tt> controls only exclusive use of each file system object and it does not that prevent re-entering the low level disk functions. For example, only <tt>disk_status</tt> function is re-entered at single volume system and every disk function is re-entered at multiple volume system. Thus the low level disk I/O layer must be always thread-safe when FatFs API is re-entered by two or more tasks.</p>
<p>FatFs module does not support the read/write collision control of duplicated open to a file. The duplicated open is permitted only when each of open method to a file is read mode. The duplicated open with one or more write mode to a file is always prohibited, and also open file must not be renamed or deleted. A violation of these rules can cause data colluption.</p>
<p>The file lock control can be enabled by <tt><ahref="config.html#fs_lock">_FS_LOCK</a></tt> option. The value of option defines the number of open objects to manage simultaneously. In this case, if any open, rename or remove that violating the file shareing rule that described above is attempted, the file function will rejected with <tt>FR_LOCKED</tt>. If number of open objects, files and sub-directories, is equal to <tt>_FS_LOCK</tt>, an extra <tt>f_open/f_opendir</tt> function will fail with <tt>FR_TOO_MANY_OPEN_FILES</tt>.</p>
<p>For good read/write throughput on the small embedded systems with limited size of memory, application programmer should consider what process is done in the FatFs module. The file data on the volume is transferred in following sequence by <tt>f_read</tt> function.</p>
<p>The file I/O buffer is a sector buffer to read/write a part of data on the sector. The sector buffer is either file private sector buffer on each file object or shared sector buffer in the file system object. The buffer configuration option <tt><ahref="config.html#fs_tiny">_FS_TINY</a></tt> determins which sector buffer is used for the file data transfer. When tiny buffer configuration (1) is selected, data memory consumption is reduced <tt>_MAX_SS</tt> bytes each file object. In this case, FatFs module uses only a sector buffer in the file system object for file data transfer and FAT/directory access. The disadvantage of the tiny buffer configuration is: the FAT data cached in the sector buffer will be lost by file data transfer and it must be reloaded at every cluster boundary. However it will be suitable for most application from view point of the decent performance and low memory comsumption.</p>
<p>Figure 1 shows that a partial sector, sector misaligned part of the file, is transferred via the file I/O buffer. At long data transfer shown in Figure 2, middle of transfer data that covers one or more sector is transferred to the application buffer directly. Figure 3 shows that the case of entier transfer data is aligned to the sector boundary. In this case, file I/O buffer is not used. On the direct transfer, the maximum extent of sectors are read with <tt>disk_read</tt> function at a time but the multiple sector transfer is divided at cluster boundary even if it is contiguous.</p>
<p>Therefore taking effort to sector aligned read/write accesss eliminates buffered data transfer and the read/write performance will be improved. Besides the effect, cached FAT data will not be flushed by file data transfer at the tiny configuration, so that it can achieve same performance as non-tiny configuration with small memory footprint.</p>
<p>To maximize the write performance of flash memory media, such as SDC, CFC and U Disk, it must be controlled in consideration of its characteristitcs.</p>
<h4>Using Mutiple-Sector Write</h4>
<divclass="rset">
Figure 6. Comparison between Multiple/Single Sector Write<br>
<p>The write throughput of the flash memory media becomes the worst at single sector write transaction. The write throughput increases as the number of sectors per a write transaction as shown in Figure 6. This effect more appers at faster interface speed and the performance ratio often becomes grater than ten. <ahref="../res/rwtest2.png">This graph</a> is clearly explaining how fast is multiple block write (W:16K, 32 sectors) than single block write (W:100, 1 sector), and also larger card tends to be slow at single block write. Number of write transactions also affects life time of the flash memory media. When compared at same amount of write data, the single sector write in Figure 6 above wears flash memory media 16 times more than multiple sector write in Figure 6 below. Single sector write is pretty pain for the flash memory media.</p>
<p>Therefore the application program should write the data in large block as possible. The ideal write chunk size and alighment is size of sector, and size of cluster is the best. Of course all layers between the application and the storage device must have consideration on multiple sector write, however most of open-source memory card drivers lack it. Do not split a multiple sector write request into single sector write transactions or the write throughput gets poor. Note that FatFs module and its sample disk drivers supprt multiple sector read/write operation. </p>
<p>When remove a file with <tt>f_unlink</tt> function, the data clusters occupied by the file are marked 'free' on the FAT. But the data sectors containing the file data are not that applied any process, so that the file data left occupies a part of the flash memory array as 'live block'. If the file data can be erased on removing the file, those data blocks will be turned into the free block pool. This may skip internal block erase operation to the data block on next write operation. As the result the write performance might be improved. FatFs can manage this function by setting <tt><ahref="config.html#use_trim">_USE_TRIM</a></tt> to 1. Note that this is an expectation of internal process of the storage device and not that always effective. Most applications will not need this function. Also <tt>f_unlink</tt> function can take a time when remove a large file.</p>
<p>If a write operation to the FAT volume is interrupted due to an accidental failure, such as sudden blackout, incorrect media removal and unrecoverable disk error, the FAT structure on the volume can be broken. Following images shows the critical section of the FatFs module.</p>
<p>An interruption in the red section can cause a cross link; as a result, the object being changed can be lost. If an interruption in the yellow section is occured, there is one or more possibility listed below.</p>
<ul>
<li>The file data being rewrited is collapsed.</li>
<li>The file being appended returns initial state.</li>
<li>The file created as new is gone.</li>
<li>The file created as new or overwritten remains but no content.</li>
<li>Efficiency of disk use gets worse due to lost clusters.</li>
<p>Each case does not affect any file not opened in write mode. To minimize risk of data loss, the critical section can be minimized by minimizing the time that file is opened in write mode or using <tt>f_sync</tt> function as shown in Figure 5.</p>
<p>FatFs has being developped as a personal project of the author, ChaN. It is free from the code anyone else wrote at current release. Following code block shows a copy of the FatFs license document that included in the source files.</p>
<p>Therefore FatFs license is one of the BSD-style licenses but there is a significant feature. FatFs is mainly intended for embedded systems. In order to extend the usability for commercial products, the redistributions of FatFs in binary form, such as embedded code, binary library and any forms without source code, does not need to include about FatFs in the documentations. This is equivalent to the 1-clause BSD license. Of course FatFs is compatible with the most of open source software licenses including GNU GPL. When you redistribute the FatFs source code with any changes or create a fork, the license can also be changed to GNU GPL, BSD-style license or any open source software license that not conflict with FatFs license.</p>
