Github/fatfs
1
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

FatFs R0.14, released October 14, 2019

Browse Source
This commit is contained in:
Christopher Williams
2019-11-14 17:35:48 -07:00
parent e76defd75c
commit 258d0cbbf0
39 changed files with 1132 additions and 672 deletions
Download Patch File Download Diff File Expand all files Collapse all files
documents/doc/appnote.html
+61 -21
View File
@@ -19,6 +19,7 @@
<li><a href="#lfn">Long File Name</a></li>
<li><a href="#unicode">Unicode API</a></li>
<li><a href="#exfat">exFAT Filesystem</a></li>
<li><a href="#lba64">64-bit LBA</a></li>
<li><a href="#reentrant">Re-entrancy</a></li>
<li><a href="#dup">Duplicated File Access</a></li>
<li><a href="#fs1">Performance Effective File Access</a></li>
@@ -38,14 +39,27 @@
The FatFs module is a middleware written in ANSI C (C89). There is no platform dependence, so long as the compiler is in compliance with C89 or later. Only exFAT feature requires C99.</li>
<li>Size of integer types<br>
<ul>
<li>When the C standard is in C89, size of <tt>char</tt>/<tt>short</tt>/<tt>long</tt> must be 8/16/32-bit respectively.</li>
<li>When it is in C99 or later, <tt>stdint.h</tt> is used to obtain the correct sizes.</li>
<li>Size of <tt>char</tt> must be 8-bit.</li>
<li>Size of <tt>int</tt>, as well as integer promotion, must be 16-bit or 32-bit.</li>
<li>When the C standard is in C89, size of <tt>short</tt> and <tt>long</tt> must be 16-bit and 32-bit respectively.</li>
<li>When it is in C99 or later, <tt>stdint.h</tt> is used to obtain the integer sizes.</li>
</ul>
Integer types used in FatFs are defined in <tt>fftypes.h</tt>. This will not be a problem on most platform. When a conflict with existing definitions occured, you must resolve it with care.</li>
</ul>
<h4>Integer Types in FatFs API</h4>
<p>Integer types used in FatFs are defined in <tt>ff.h</tt> as described below. It is based on Win32 API (<tt>windef.h</tt>). This will not be a problem on most platform. When a conflict with existing definitions occured, you must resolve it with care.</p>
<dl>
<dt><tt>BYTE</tt></dt><dd>8-bit unsigned integer in range of 0 to 2<sup>8</sup> - 1.</dd>
<dt><tt>WORD</tt></dt><dd>16-bit unsigned integer in range of 0 to 2<sup>16</sup> - 1.</dd>
<dt><tt>DWORD</tt></dt><dd>32-bit unsigned integer in range of 0 to 2<sup>32</sup> - 1.</dd>
<dt><tt>QWORD</tt></dt><dd>64-bit unsigned integer in range of 0 to 2<sup>64</sup> - 1.</dd>
<dt><tt>UINT</tt></dt><dd>Alias of <tt>unsigned int</tt> used to specify any number.</dd>
<dt><tt>WCHAR</tt></dt><dd>Alias of <tt>WORD</tt> used to specify a UTF-16 character unit.</dd>
<dt><tt>TCHAR</tt></dt><dd>Alias of <tt>char</tt>, <tt>WCHAR</tt> or <tt>DWORD</tt> used to specify a character unit.</dd>
<dt><tt>FSIZE_t</tt></dt><dd>Alias of <tt>DWORD</tt> or <tt>QWORD</tt> used to address file offset and to specify file size.</dd>
<dt><tt>LBA_t</tt></dt><dd>Alias of <tt>DWORD</tt> or <tt>QWORD</tt> used to address sectors in LBA and to specify number of sectors.</dd>
</dl>
<h4>System Organizations</h4>
<p>The dependency diagram shown below is a typical, but not specific, configuration of the embedded system with FatFs module.</p>
<p><img src="../res/modules.png" width="580" height="280" alt="dependency diagram"></p>
@@ -73,11 +87,12 @@ Integer types used in FatFs are defined in <tt>fftypes.h</tt>. This will not be
<ul>
<li>Filesystem type: FAT, FAT32(rev0.0) and exFAT(rev1.0).</li>
<li>Number of open files: Unlimited. (depends on available memory)</li>
<li>Number of volumes: Upto 10.</li>
<li>Volume size: Upto 2 TB at 512 bytes/sector.</li>
<li>File size: Upto 4 GB - 1 on FAT volume and virtually unlimited on exFAT volume.</li>
<li>Cluster size: Upto 128 sectors on FAT volume and upto 16 MB on exFAT volume.</li>
<li>Number of volumes: Up to 10.</li>
<li>Sector size: 512, 1024, 2048 and 4096 bytes.</li>
<li>Minimum volume size: 128 sectors.</li>
<li>Maximum volume size: 2<sup>32</sup> - 1 sectors in 32-bit LBA, virtually unlimited in 64-bit LBA with exFAT.</li>
<li>Maximum file size: 2<sup>32</sup> - 1 bytes on FAT volume, virtually unlimited on exFAT volume.</li>
<li>Cluster size: Upto 128 sectors on FAT volume and up to 16 MB on exFAT volume.</li>
</ul>
</div>
@@ -107,7 +122,7 @@ And any other options are left unchanged from original setting.
</div>
<div class="para doc" id="reduce">
<h3>Reducing Modle Size</h3>
<h3>Reducing Module Size</h3>
<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>
<table class="lst2">
<tr><td rowspan="2">Function</td><td colspan="4">FF_FS_<br>MINIMIZE</td><td colspan="2">FF_FS_<br>READONLY</td><td colspan="2">FF_USE_<br>STRFUNC</td><td colspan="3">FF_FS_<br>RPATH</td><td colspan="2">FF_USE_<br>FIND</td><td colspan="2">FF_USE_<br>CHMOD</td><td colspan="2">FF_USE_<br>EXPAND</td><td colspan="2">FF_USE_<br>LABEL</td><td colspan="2">FF_USE_<br>MKFS</td><td colspan="2">FF_USE_<br>FORWARD</td><td colspan="2">FF_MULTI_<br>PARTITION</td></tr>
@@ -151,7 +166,8 @@ And any other options are left unchanged from original setting.
<div class="para doc" id="lfn">
<h3>Long File Name</h3>
<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><a href="config.html#use_lfn">FF_USE_LFN</a></tt> to 1, 2 or 3, and add <tt>ffunicode.c</tt> to the project. The LFN requiers a certain working buffer in addition. The buffer size can be configured by <tt><a href="config.html#max_lfn">FF_MAX_LFN</a></tt> according to the available memory. The length of an LFN will be up to 255 characters, so that the <tt>FF_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>FF_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>(FF_MAX_LFN + 1) * 2</tt> bytes and additional <tt>(FF_MAX_LFN + 44) / 15 * 32</tt> bytes when exFAT is enabled.</p>
<p>FatFs module supports the long file name (LFN) extension of the FAT filesystem. The two different file names, short file name (SFN) and LFN, of a file is transparent on the API. The support for LFN feature is disabled by default. To enable the LFN, set <tt><a href="config.html#use_lfn">FF_USE_LFN</a></tt> to 1, 2 or 3, and add <tt>ffunicode.c</tt> to the project. The LFN feature requiers a certain working buffer. The buffer size can be configured by <tt><a href="config.html#max_lfn">FF_MAX_LFN</a></tt> according to the available memory. The length of an LFN can be up to 255 characters, so that the <tt>FF_MAX_LFN</tt> should be set to 255 for all existing 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 feature, <tt>FF_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 LFN working buffer occupies <tt>(FF_MAX_LFN + 1) * 2</tt> bytes and additional <tt>(FF_MAX_LFN + 44) / 15 * 32</tt> bytes when exFAT is enabled.</p>
<h4>Impact upon Module Size</h4>
<table class="lst2 rset">
<caption>With LFN at CM3 + gcc</caption>
<tr><th><tt>FF_CODE_PAGE</tt></th><th>Code size</th></tr>
@@ -162,9 +178,9 @@ And any other options are left unchanged from original setting.
<tr><td>950 (Traditional Chinese)</td><td>+111k</td></tr>
<tr><td>0 (All code pages)</td><td>+486k</td></tr>
</table>
<p>When the LFN is enabled, the module size will be increased depends on the configured code page. Right table shows increment of code size by LFN function at different 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.</p>
<p>As the result, the FatFs with LFN enebled with DBCS code pages will not able to be ported on the most 8-bit MCU systems. If the target system is in legacy-free, in only Unicode and any ANSI/OEM code is not used at all, the code page setting gets meaningless. You will able to reduce the code size by configureing FatFs for Unicode with any SBCS code page.</p>
<p>There ware some restrictions on using LFN for open source project because the support for LFN on the FAT volume was a patent of Microsoft Corporation. The related patents have expired and using the LFN function has got free for any projects.</p>
<p>When the LFN is enabled, the module size will be increased depends on the configured code page. Right table shows the increment of code size in 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.</p>
<p>As the result, the FatFs with LFN enabled with DBCS code pages will not able to be ported on the most 8-bit MCU systems. If the target system is in legacy-free, in only Unicode and any ANSI/OEM code is not used at all, the code page setting gets meaningless. You will able to reduce the code size by configureing FatFs for Unicode with any SBCS code page.</p>
<p>There ware some restrictions on using LFN for open source project because the LFN extension on the FAT filesystem was a patent of Microsoft Corporation. But the related patents all have expired and using the LFN feature has got free for any projects.</p>
</div>
<div class="para doc" id="unicode">
@@ -175,17 +191,40 @@ And any other options are left unchanged from original setting.
<div class="para doc" id="exfat">
<h3>exFAT Filesystem</h3>
<p>The exFAT (Microsoft's Extended File Allocation Table) filesystem is a succession of the FAT/FAT32 filesystem which has been widely used in embedded systems, consumer devices and portable storage media. It is adopted by SDA (SD Association) as the regular filesystem for SDXC card, 64 GB and larger, and they are being shipped with this format. Therefore the exFAT is one of the standard filesystems for removable media as well as FAT. The exFAT filesystem allows the file size beyond the 4 GB limit what FAT filesystem allows upto and some filesystem overhead, especially cluster allocation delay, are reduced as well. These features allow to record the large media file without dividing into some files and improve the write throughput to the file.</p>
<p>Note that the exFAT is a patent of Microsoft Corporation. The support for exFAT of FatFs is an implementation based on <cite>US. Pat. App. Pub. No. 2009/0164440 A1</cite>. FatFs module can switch the exFAT on or off by configuration option, <tt><a href="config.html#fs_exfat">FF_FS_EXFAT</a></tt>. When enable the exFAT on the commercial products, a license by Microsoft will be needed depends on the final destination of the products.</p>
<p><em>Remark: Enabling exFAT discards C89 compatibility because of need for 64-bit integer type.</em></p>
<p>The exFAT (Microsoft's Extended File Allocation Table) filesystem is a succession of the FAT/FAT32 filesystem which has been widely used in embedded systems, consumer devices and portable storage media. It is adopted by SDA (SD Association) as the filesystem for SDXC card, 64 GB and larger, and they are being shipped with this format. Therefore the exFAT is one of the standard filesystems for removable media as well as FAT. The exFAT filesystem allows the file size beyond the 4 GB limit what FAT filesystem allows up to and some filesystem overhead, especially cluster allocation delay, are reduced as well. These features allow to record the large data without dividing into some files and improve the write throughput to the file.</p>
<p>Note that the exFAT filesystem is a patent of Microsoft Corporation. The exFAT feature of FatFs is an implementation based on <cite>US. Pat. App. Pub. No. 2009/0164440 A1</cite>. FatFs module can switch the exFAT on or off by a configuration option, <tt><a href="config.html#fs_exfat">FF_FS_EXFAT</a></tt>. When enable the exFAT for the commercial products, a license by Microsoft will be needed depends on the final destination of the products.</p>
<p><em>Remarks: Enabling exFAT discards C89 compatibility and it wants C99 because of need for 64-bit integer type.</em></p>
</div>
<div class="para doc" id="lba64">
<h3>64-bit LBA</h3>
<p>LBA (Logical Block Addressing) is an addressing method to specify the location of data block, called sector, on the storage media. It is a simple linear address beginning from 0 as the first block of data. The host system does not need to consider how the data block is located and managed in the storage device. FatFs supports only LBA for the media access. 32-bit LBA is a common size at the interface of the most storage devices. It can address up to 2<sup>32</sup> blocks, 2 TB in 512 bytes/sector. When a storage device beyond 2 TB is used, increasing sector size or 64-bit LBA will be needed to address the entire block of the storage device.</p>
<p>By default, FatFs uses 32-bit LBA for media access interface. FatFs can also switch it to 64-bit LBA by a configuration option <tt><a href="config.html#fs_lba64">FF_LBA64</a></tt>. It also enables GPT (GUID Partition Table) for partiotion management on the storage media. For further information about GPT, refer to <tt><a href="mkfs.html">f_mkfs</a></tt> and <tt><a href="fdisk.html">f_fdisk</a></tt> function.</p>
</div>
<div class="para doc" id="reentrant">
<h3>Re-entrancy</h3>
<p>The file operations to the <em>different volumes</em> each other is always re-entrant regardless of configurations except when LFN is enabled with static working buffer. It can work concurrently without any mutual exclusion.</p>
<p>The file operations to the <em>same volume</em> is not re-entrant. It can also be configured for thread-safe by option <tt><a href="config.html#fs_reentrant">FF_FS_REENTRANT</a></tt>. In this case, also the OS dependent synchronization control functions, <tt>ff_cre_syncobj/ff_del_syncobj/ff_req_grant/ff_rel_grant</tt>, need to be added to the project. There are some examples in the <tt>ffsystem.c</tt>. When a file function is called while the volume is being accessed by another 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>FF_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>The file operations of two tasks to the <em>different volumes</em> each other is always re-entrant regardless of the configurations except when LFN is enabled with static working buffer (<tt>FF_USE_LFN = 1</tt>). It can work concurrently without any mutual exclusion.</p>
<p>The file operations of two tasks to the <em>same volume</em> is not re-entrant in default. FatFs can also be configured to make it thread-safe by option <tt><a href="config.html#fs_reentrant">FF_FS_REENTRANT</a></tt>. In this case, also the OS dependent synchronization control functions, <tt>ff_cre_syncobj/ff_del_syncobj/ff_req_grant/ff_rel_grant</tt>, need to be added to the project. There are some examples in the <tt>ffsystem.c</tt>. When a file function is called while the volume is being accessed by another 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>FF_TIMEOUT</tt>, the file function will abort with <tt>FR_TIMEOUT</tt>. The timeout feature might not be supported on the some RTOSs.</p>
<p>There is an exception on the re-entrancy for <tt>f_mount/f_mkfs</tt> function. These volume management functions are not re-entrant to the same volume. When use these functions, other tasks need to avoid to access the volume.</p>
<p>Remarks: This section describes on the re-entrancy of the FatFs module itself. The <tt>FF_FS_REENTRANT</tt> option enables only exclusive use of each filesystem objects and FatFs never prevents to re-enter <tt>disk_*()</tt> functions. Thus the low level disk I/O layer must be always thread-safe when FatFs API is re-entered for different volumes.</p>
<div class="rset">
<table class="lst2">
<tr><th><tt>Function</tt></th><th>Case 1</th><th>Case 2</th><th>Case 3</th></tr>
<tr><td>disk_status</td><td>Yes</td><td>Yes</td><td>Yes(*)</td></tr>
<tr><td>disk_initialize</td><td>No</td><td>Yes</td><td>Yes(*)</td></tr>
<tr><td>disk_read</td><td>No</td><td>Yes</td><td>Yes(*)</td></tr>
<tr><td>disk_write</td><td>No</td><td>Yes</td><td>Yes(*)</td></tr>
<tr><td>disk_ioctl</td><td>No</td><td>Yes</td><td>Yes(*)</td></tr>
<tr><td>get_fattime</td><td>No</td><td>Yes</td><td>Yes</td></tr>
</table>
<small>
Case 1: Same volume.<br>
Case 2: Different volume on the same drive.<br>
Case 3: Different volume on the different drive.<br>
(*) In only different drive number.
</small>
</div>
<p>Remarks: This section describes on the re-entrancy of the FatFs module itself. The <tt>FF_FS_REENTRANT</tt> option enables only exclusive use of each filesystem objects and FatFs does not that prevent to re-enter the low level disk functions. Thus the low level disk I/O layer needs to be always thread-safe when FatFs API is re-entered for different volumes. Right table shows which low level function can be re-entered when FatFs API is re-entered on some conditions.</p>
</div>
<div class="para doc" id="dup">
@@ -227,7 +266,7 @@ Figure 6. Comparison between Multiple/Single Sector Write<br>
<div class="para doc" id="critical">
<h3>Critical Section</h3>
<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>If a write operation to the FAT volume is interrupted due to an accidental failure, such as sudden blackout, wrong 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>
<div class="lset">
Figure 4. Long critical section<br>
<img src="../res/f4.png" width="320" height="436" alt="fig.4">
@@ -255,9 +294,10 @@ Figure 5. Minimized critical section<br>
<li><a href="http://elm-chan.org/fsw/ff/res/app1.c">Open or create a file for append</a> (for R0.12 and earlier)</li>
<li><a href="http://elm-chan.org/fsw/ff/res/app2.c">Delete a non-empty sub-directory</a> (for R0.12 and later)</li>
<li><a href="http://elm-chan.org/fsw/ff/res/app3.c">Allocate contiguous area to the file</a> (for R0.11a and earlier)</li>
<li><a href="http://elm-chan.org/fsw/ff/res/app4.c">Compatibility checker for low level disk I/O module</a></li>
<li><a href="http://elm-chan.org/fsw/ff/res/mkfatimg.zip">FAT volume image creator</a></li>
<li><a href="http://elm-chan.org/fsw/ff/res/app5.c">Test if the file is contiguous or not</a></li>
<li><a href="http://elm-chan.org/fsw/ff/res/app4.c">Compatibility checker for low level disk I/O module</a></li>
<li><a href="http://elm-chan.org/fsw/ff/res/app6.c">Performance checker for low level disk I/O module</a></li>
<li><a href="http://elm-chan.org/fsw/ff/res/mkfatimg.zip">FAT volume image creator</a></li>
</ol>
</div>