+
+/*
+ * Copyright (c) 2006-2012 by Roland Riegel <feedback@roland-riegel.de>
+ *
+ * This file is free software; you can redistribute it and/or modify
+ * it under the terms of either the GNU General Public License version 2
+ * or the GNU Lesser General Public License version 2.1, both as
+ * published by the Free Software Foundation.
+ */
+
+#include "byteordering.h"
+#include "partition.h"
+#include "fat.h"
+#include "fat_config.h"
+#include "sd-reader_config.h"
+
+#include <string.h>
+
+#if USE_DYNAMIC_MEMORY
+ #include <stdlib.h>
+#endif
+
+/**
+ * \addtogroup fat FAT support
+ *
+ * This module implements FAT16/FAT32 read and write access.
+ *
+ * The following features are supported:
+ * - File names up to 31 characters long.
+ * - Unlimited depth of subdirectories.
+ * - Short 8.3 and long filenames.
+ * - Creating and deleting files.
+ * - Reading and writing from and to files.
+ * - File resizing.
+ * - File sizes of up to 4 gigabytes.
+ *
+ * @{
+ */
+/**
+ * \file
+ * FAT implementation (license: GPLv2 or LGPLv2.1)
+ *
+ * \author Roland Riegel
+ */
+
+/**
+ * \addtogroup fat_config FAT configuration
+ * Preprocessor defines to configure the FAT implementation.
+ */
+
+/**
+ * \addtogroup fat_fs FAT access
+ * Basic functions for handling a FAT filesystem.
+ */
+
+/**
+ * \addtogroup fat_file FAT file functions
+ * Functions for managing files.
+ */
+
+/**
+ * \addtogroup fat_dir FAT directory functions
+ * Functions for managing directories.
+ */
+
+/**
+ * @}
+ */
+
+#define FAT16_CLUSTER_FREE 0x0000
+#define FAT16_CLUSTER_RESERVED_MIN 0xfff0
+#define FAT16_CLUSTER_RESERVED_MAX 0xfff6
+#define FAT16_CLUSTER_BAD 0xfff7
+#define FAT16_CLUSTER_LAST_MIN 0xfff8
+#define FAT16_CLUSTER_LAST_MAX 0xffff
+
+#define FAT32_CLUSTER_FREE 0x00000000
+#define FAT32_CLUSTER_RESERVED_MIN 0x0ffffff0
+#define FAT32_CLUSTER_RESERVED_MAX 0x0ffffff6
+#define FAT32_CLUSTER_BAD 0x0ffffff7
+#define FAT32_CLUSTER_LAST_MIN 0x0ffffff8
+#define FAT32_CLUSTER_LAST_MAX 0x0fffffff
+
+#define FAT_DIRENTRY_DELETED 0xe5
+#define FAT_DIRENTRY_LFNLAST (1 << 6)
+#define FAT_DIRENTRY_LFNSEQMASK ((1 << 6) - 1)
+
+/* Each entry within the directory table has a size of 32 bytes
+ * and either contains a 8.3 DOS-style file name or a part of a
+ * long file name, which may consist of several directory table
+ * entries at once.
+ *
+ * multi-byte integer values are stored little-endian!
+ *
+ * 8.3 file name entry:
+ * ====================
+ * offset length description
+ * 0 8 name (space padded)
+ * 8 3 extension (space padded)
+ * 11 1 attributes (FAT_ATTRIB_*)
+ *
+ * long file name (lfn) entry ordering for a single file name:
+ * ===========================================================
+ * LFN entry n
+ * ...
+ * LFN entry 2
+ * LFN entry 1
+ * 8.3 entry (see above)
+ *
+ * lfn entry:
+ * ==========
+ * offset length description
+ * 0 1 ordinal field
+ * 1 2 unicode character 1
+ * 3 3 unicode character 2
+ * 5 3 unicode character 3
+ * 7 3 unicode character 4
+ * 9 3 unicode character 5
+ * 11 1 attribute (always 0x0f)
+ * 12 1 type (reserved, always 0)
+ * 13 1 checksum
+ * 14 2 unicode character 6
+ * 16 2 unicode character 7
+ * 18 2 unicode character 8
+ * 20 2 unicode character 9
+ * 22 2 unicode character 10
+ * 24 2 unicode character 11
+ * 26 2 cluster (unused, always 0)
+ * 28 2 unicode character 12
+ * 30 2 unicode character 13
+ *
+ * The ordinal field contains a descending number, from n to 1.
+ * For the n'th lfn entry the ordinal field is or'ed with 0x40.
+ * For deleted lfn entries, the ordinal field is set to 0xe5.
+ */
+
+struct fat_header_struct
+{
+ offset_t size;
+
+ offset_t fat_offset;
+ uint32_t fat_size;
+
+ uint16_t sector_size;
+ uint16_t cluster_size;
+
+ offset_t cluster_zero_offset;
+
+ offset_t root_dir_offset;
+#if FAT_FAT32_SUPPORT
+ cluster_t root_dir_cluster;
+#endif
+};
+
+struct fat_fs_struct
+{
+ struct partition_struct* partition;
+ struct fat_header_struct header;
+ cluster_t cluster_free;
+};
+
+struct fat_file_struct
+{
+ struct fat_fs_struct* fs;
+ struct fat_dir_entry_struct dir_entry;
+ offset_t pos;
+ cluster_t pos_cluster;
+};
+
+struct fat_dir_struct
+{
+ struct fat_fs_struct* fs;
+ struct fat_dir_entry_struct dir_entry;
+ cluster_t entry_cluster;
+ uint16_t entry_offset;
+};
+
+struct fat_read_dir_callback_arg
+{
+ struct fat_dir_entry_struct* dir_entry;
+ uintptr_t bytes_read;
+#if FAT_LFN_SUPPORT
+ uint8_t checksum;
+#endif
+ uint8_t finished;
+};
+
+struct fat_usage_count_callback_arg
+{
+ cluster_t cluster_count;
+ uintptr_t buffer_size;
+};
+
+#if !USE_DYNAMIC_MEMORY
+static struct fat_fs_struct fat_fs_handles[FAT_FS_COUNT];
+static struct fat_file_struct fat_file_handles[FAT_FILE_COUNT];
+static struct fat_dir_struct fat_dir_handles[FAT_DIR_COUNT];
+#endif
+
+static uint8_t fat_read_header(struct fat_fs_struct* fs);
+static cluster_t fat_get_next_cluster(const struct fat_fs_struct* fs, cluster_t cluster_num);
+static offset_t fat_cluster_offset(const struct fat_fs_struct* fs, cluster_t cluster_num);
+static uint8_t fat_dir_entry_read_callback(uint8_t* buffer, offset_t offset, void* p);
+#if FAT_LFN_SUPPORT
+static uint8_t fat_calc_83_checksum(const uint8_t* file_name_83);
+#endif
+
+static uint8_t fat_get_fs_free_16_callback(uint8_t* buffer, offset_t offset, void* p);
+#if FAT_FAT32_SUPPORT
+static uint8_t fat_get_fs_free_32_callback(uint8_t* buffer, offset_t offset, void* p);
+#endif
+
+#if FAT_WRITE_SUPPORT
+static cluster_t fat_append_clusters(struct fat_fs_struct* fs, cluster_t cluster_num, cluster_t count);
+static uint8_t fat_free_clusters(struct fat_fs_struct* fs, cluster_t cluster_num);
+static uint8_t fat_terminate_clusters(struct fat_fs_struct* fs, cluster_t cluster_num);
+static uint8_t fat_clear_cluster(const struct fat_fs_struct* fs, cluster_t cluster_num);
+static uintptr_t fat_clear_cluster_callback(uint8_t* buffer, offset_t offset, void* p);
+static offset_t fat_find_offset_for_dir_entry(struct fat_fs_struct* fs, const struct fat_dir_struct* parent, const struct fat_dir_entry_struct* dir_entry);
+static uint8_t fat_write_dir_entry(const struct fat_fs_struct* fs, struct fat_dir_entry_struct* dir_entry);
+#if FAT_DATETIME_SUPPORT
+static void fat_set_file_modification_date(struct fat_dir_entry_struct* dir_entry, uint16_t year, uint8_t month, uint8_t day);
+static void fat_set_file_modification_time(struct fat_dir_entry_struct* dir_entry, uint8_t hour, uint8_t min, uint8_t sec);
+#endif
+#endif
+
+/**
+ * \ingroup fat_fs
+ * Opens a FAT filesystem.
+ *
+ * \param[in] partition Discriptor of partition on which the filesystem resides.
+ * \returns 0 on error, a FAT filesystem descriptor on success.
+ * \see fat_close
+ */
+struct fat_fs_struct* fat_open(struct partition_struct* partition)
+{
+ if(!partition ||
+#if FAT_WRITE_SUPPORT
+ !partition->device_write ||
+ !partition->device_write_interval
+#else
+ 0
+#endif
+ )
+ return 0;
+
+#if USE_DYNAMIC_MEMORY
+ struct fat_fs_struct* fs = malloc(sizeof(*fs));
+ if(!fs)
+ return 0;
+#else
+ struct fat_fs_struct* fs = fat_fs_handles;
+ uint8_t i;
+ for(i = 0; i < FAT_FS_COUNT; ++i)
+ {
+ if(!fs->partition)
+ break;
+
+ ++fs;
+ }
+ if(i >= FAT_FS_COUNT)
+ return 0;
+#endif
+
+ memset(fs, 0, sizeof(*fs));
+
+ fs->partition = partition;
+ if(!fat_read_header(fs))
+ {
+#if USE_DYNAMIC_MEMORY
+ free(fs);
+#else
+ fs->partition = 0;
+#endif
+ return 0;
+ }
+
+ return fs;
+}
+
+/**
+ * \ingroup fat_fs
+ * Closes a FAT filesystem.
+ *
+ * When this function returns, the given filesystem descriptor
+ * will be invalid.
+ *
+ * \param[in] fs The filesystem to close.
+ * \see fat_open
+ */
+void fat_close(struct fat_fs_struct* fs)
+{
+ if(!fs)
+ return;
+
+#if USE_DYNAMIC_MEMORY
+ free(fs);
+#else
+ fs->partition = 0;
+#endif
+}
+
+/**
+ * \ingroup fat_fs
+ * Reads and parses the header of a FAT filesystem.
+ *
+ * \param[in,out] fs The filesystem for which to parse the header.
+ * \returns 0 on failure, 1 on success.
+ */
+uint8_t fat_read_header(struct fat_fs_struct* fs)
+{
+ if(!fs)
+ return 0;
+
+ struct partition_struct* partition = fs->partition;
+ if(!partition)
+ return 0;
+
+ /* read fat parameters */
+#if FAT_FAT32_SUPPORT
+ uint8_t buffer[37];
+#else
+ uint8_t buffer[25];
+#endif
+ offset_t partition_offset = (offset_t) partition->offset * 512;
+ if(!partition->device_read(partition_offset + 0x0b, buffer, sizeof(buffer)))
+ return 0;
+
+ uint16_t bytes_per_sector = read16(&buffer[0x00]);
+ uint16_t reserved_sectors = read16(&buffer[0x03]);
+ uint8_t sectors_per_cluster = buffer[0x02];
+ uint8_t fat_copies = buffer[0x05];
+ uint16_t max_root_entries = read16(&buffer[0x06]);
+ uint16_t sector_count_16 = read16(&buffer[0x08]);
+ uint16_t sectors_per_fat = read16(&buffer[0x0b]);
+ uint32_t sector_count = read32(&buffer[0x15]);
+#if FAT_FAT32_SUPPORT
+ uint32_t sectors_per_fat32 = read32(&buffer[0x19]);
+ uint32_t cluster_root_dir = read32(&buffer[0x21]);
+#endif
+
+ if(sector_count == 0)
+ {
+ if(sector_count_16 == 0)
+ /* illegal volume size */
+ return 0;
+ else
+ sector_count = sector_count_16;
+ }
+#if FAT_FAT32_SUPPORT
+ if(sectors_per_fat != 0)
+ sectors_per_fat32 = sectors_per_fat;
+ else if(sectors_per_fat32 == 0)
+ /* this is neither FAT16 nor FAT32 */
+ return 0;
+#else
+ if(sectors_per_fat == 0)
+ /* this is not a FAT16 */
+ return 0;
+#endif
+
+ /* determine the type of FAT we have here */
+ uint32_t data_sector_count = sector_count
+ - reserved_sectors
+#if FAT_FAT32_SUPPORT
+ - sectors_per_fat32 * fat_copies
+#else
+ - (uint32_t) sectors_per_fat * fat_copies
+#endif
+ - ((max_root_entries * 32 + bytes_per_sector - 1) / bytes_per_sector);
+ uint32_t data_cluster_count = data_sector_count / sectors_per_cluster;
+ if(data_cluster_count < 4085)
+ /* this is a FAT12, not supported */
+ return 0;
+ else if(data_cluster_count < 65525)
+ /* this is a FAT16 */
+ partition->type = PARTITION_TYPE_FAT16;
+ else
+ /* this is a FAT32 */
+ partition->type = PARTITION_TYPE_FAT32;
+
+ /* fill header information */
+ struct fat_header_struct* header = &fs->header;
+ memset(header, 0, sizeof(*header));
+
+ header->size = (offset_t) sector_count * bytes_per_sector;
+
+ header->fat_offset = /* jump to partition */
+ partition_offset +
+ /* jump to fat */
+ (offset_t) reserved_sectors * bytes_per_sector;
+ header->fat_size = (data_cluster_count + 2) * (partition->type == PARTITION_TYPE_FAT16 ? 2 : 4);
+
+ header->sector_size = bytes_per_sector;
+ header->cluster_size = (uint16_t) bytes_per_sector * sectors_per_cluster;
+
+#if FAT_FAT32_SUPPORT
+ if(partition->type == PARTITION_TYPE_FAT16)
+#endif
+ {
+ header->root_dir_offset = /* jump to fats */
+ header->fat_offset +
+ /* jump to root directory entries */
+ (offset_t) fat_copies * sectors_per_fat * bytes_per_sector;
+
+ header->cluster_zero_offset = /* jump to root directory entries */
+ header->root_dir_offset +
+ /* skip root directory entries */
+ (offset_t) max_root_entries * 32;
+ }
+#if FAT_FAT32_SUPPORT
+ else
+ {
+ header->cluster_zero_offset = /* jump to fats */
+ header->fat_offset +
+ /* skip fats */
+ (offset_t) fat_copies * sectors_per_fat32 * bytes_per_sector;
+
+ header->root_dir_cluster = cluster_root_dir;
+ }
+#endif
+
+ return 1;
+}
+
+/**
+ * \ingroup fat_fs
+ * Retrieves the next following cluster of a given cluster.
+ *
+ * Using the filesystem file allocation table, this function returns
+ * the number of the cluster containing the data directly following
+ * the data within the cluster with the given number.
+ *
+ * \param[in] fs The filesystem for which to determine the next cluster.
+ * \param[in] cluster_num The number of the cluster for which to determine its successor.
+ * \returns The wanted cluster number, or 0 on error.
+ */
+cluster_t fat_get_next_cluster(const struct fat_fs_struct* fs, cluster_t cluster_num)
+{
+ if(!fs || cluster_num < 2)
+ return 0;
+
+#if FAT_FAT32_SUPPORT
+ if(fs->partition->type == PARTITION_TYPE_FAT32)
+ {
+ /* read appropriate fat entry */
+ uint32_t fat_entry;
+ if(!fs->partition->device_read(fs->header.fat_offset + (offset_t) cluster_num * sizeof(fat_entry), (uint8_t*) &fat_entry, sizeof(fat_entry)))
+ return 0;
+
+ /* determine next cluster from fat */
+ cluster_num = ltoh32(fat_entry);
+
+ if(cluster_num == FAT32_CLUSTER_FREE ||
+ cluster_num == FAT32_CLUSTER_BAD ||
+ (cluster_num >= FAT32_CLUSTER_RESERVED_MIN && cluster_num <= FAT32_CLUSTER_RESERVED_MAX) ||
+ (cluster_num >= FAT32_CLUSTER_LAST_MIN && cluster_num <= FAT32_CLUSTER_LAST_MAX))
+ return 0;
+ }
+ else
+#endif
+ {
+ /* read appropriate fat entry */
+ uint16_t fat_entry;
+ if(!fs->partition->device_read(fs->header.fat_offset + (offset_t) cluster_num * sizeof(fat_entry), (uint8_t*) &fat_entry, sizeof(fat_entry)))
+ return 0;
+
+ /* determine next cluster from fat */
+ cluster_num = ltoh16(fat_entry);
+
+ if(cluster_num == FAT16_CLUSTER_FREE ||
+ cluster_num == FAT16_CLUSTER_BAD ||
+ (cluster_num >= FAT16_CLUSTER_RESERVED_MIN && cluster_num <= FAT16_CLUSTER_RESERVED_MAX) ||
+ (cluster_num >= FAT16_CLUSTER_LAST_MIN && cluster_num <= FAT16_CLUSTER_LAST_MAX))
+ return 0;
+ }
+
+ return cluster_num;
+}
+
+#if DOXYGEN || FAT_WRITE_SUPPORT
+/**
+ * \ingroup fat_fs
+ * Appends a new cluster chain to an existing one.
+ *
+ * Set cluster_num to zero to create a completely new one.
+ *
+ * \param[in] fs The file system on which to operate.
+ * \param[in] cluster_num The cluster to which to append the new chain.
+ * \param[in] count The number of clusters to allocate.
+ * \returns 0 on failure, the number of the first new cluster on success.
+ */
+cluster_t fat_append_clusters(struct fat_fs_struct* fs, cluster_t cluster_num, cluster_t count)
+{
+ if(!fs)
+ return 0;
+
+ device_read_t device_read = fs->partition->device_read;
+ device_write_t device_write = fs->partition->device_write;
+ offset_t fat_offset = fs->header.fat_offset;
+ cluster_t count_left = count;
+ cluster_t cluster_current = fs->cluster_free;
+ cluster_t cluster_next = 0;
+ cluster_t cluster_count;
+ uint16_t fat_entry16;
+#if FAT_FAT32_SUPPORT
+ uint32_t fat_entry32;
+ uint8_t is_fat32 = (fs->partition->type == PARTITION_TYPE_FAT32);
+
+ if(is_fat32)
+ cluster_count = fs->header.fat_size / sizeof(fat_entry32);
+ else
+#endif
+ cluster_count = fs->header.fat_size / sizeof(fat_entry16);
+
+ fs->cluster_free = 0;
+ for(cluster_t cluster_left = cluster_count; cluster_left > 0; --cluster_left, ++cluster_current)
+ {
+ if(cluster_current < 2 || cluster_current >= cluster_count)
+ cluster_current = 2;
+
+#if FAT_FAT32_SUPPORT
+ if(is_fat32)
+ {
+ if(!device_read(fat_offset + (offset_t) cluster_current * sizeof(fat_entry32), (uint8_t*) &fat_entry32, sizeof(fat_entry32)))
+ return 0;
+ }
+ else
+#endif
+ {
+ if(!device_read(fat_offset + (offset_t) cluster_current * sizeof(fat_entry16), (uint8_t*) &fat_entry16, sizeof(fat_entry16)))
+ return 0;
+ }
+
+#if FAT_FAT32_SUPPORT
+ if(is_fat32)
+ {
+ /* check if this is a free cluster */
+ if(fat_entry32 != HTOL32(FAT32_CLUSTER_FREE))
+ continue;
+
+ /* If we don't need this free cluster for the
+ * current allocation, we keep it in mind for
+ * the next time.
+ */
+ if(count_left == 0)
+ {
+ fs->cluster_free = cluster_current;
+ break;
+ }
+
+ /* allocate cluster */
+ if(cluster_next == 0)
+ fat_entry32 = HTOL32(FAT32_CLUSTER_LAST_MAX);
+ else
+ fat_entry32 = htol32(cluster_next);
+
+ if(!device_write(fat_offset + (offset_t) cluster_current * sizeof(fat_entry32), (uint8_t*) &fat_entry32, sizeof(fat_entry32)))
+ break;
+ }
+ else
+#endif
+ {
+ /* check if this is a free cluster */
+ if(fat_entry16 != HTOL16(FAT16_CLUSTER_FREE))
+ continue;
+
+ /* If we don't need this free cluster for the
+ * current allocation, we keep it in mind for
+ * the next time.
+ */
+ if(count_left == 0)
+ {
+ fs->cluster_free = cluster_current;
+ break;
+ }
+
+ /* allocate cluster */
+ if(cluster_next == 0)
+ fat_entry16 = HTOL16(FAT16_CLUSTER_LAST_MAX);
+ else
+ fat_entry16 = htol16((uint16_t) cluster_next);
+
+ if(!device_write(fat_offset + (offset_t) cluster_current * sizeof(fat_entry16), (uint8_t*) &fat_entry16, sizeof(fat_entry16)))
+ break;
+ }
+
+ cluster_next = cluster_current;
+ --count_left;
+ }
+
+ do
+ {
+ if(count_left > 0)
+ break;
+
+ /* We allocated a new cluster chain. Now join
+ * it with the existing one (if any).
+ */
+ if(cluster_num >= 2)
+ {
+#if FAT_FAT32_SUPPORT
+ if(is_fat32)
+ {
+ fat_entry32 = htol32(cluster_next);
+
+ if(!device_write(fat_offset + (offset_t) cluster_num * sizeof(fat_entry32), (uint8_t*) &fat_entry32, sizeof(fat_entry32)))
+ break;
+ }
+ else
+#endif
+ {
+ fat_entry16 = htol16((uint16_t) cluster_next);
+
+ if(!device_write(fat_offset + (offset_t) cluster_num * sizeof(fat_entry16), (uint8_t*) &fat_entry16, sizeof(fat_entry16)))
+ break;
+ }
+ }
+
+ return cluster_next;
+
+ } while(0);
+
+ /* No space left on device or writing error.
+ * Free up all clusters already allocated.
+ */
+ fat_free_clusters(fs, cluster_next);
+
+ return 0;
+}
+#endif
+
+#if DOXYGEN || FAT_WRITE_SUPPORT
+/**
+ * \ingroup fat_fs
+ * Frees a cluster chain, or a part thereof.
+ *
+ * Marks the specified cluster and all clusters which are sequentially
+ * referenced by it as free. They may then be used again for future
+ * file allocations.
+ *
+ * \note If this function is used for freeing just a part of a cluster
+ * chain, the new end of the chain is not correctly terminated
+ * within the FAT. Use fat_terminate_clusters() instead.
+ *
+ * \param[in] fs The filesystem on which to operate.
+ * \param[in] cluster_num The starting cluster of the chain which to free.
+ * \returns 0 on failure, 1 on success.
+ * \see fat_terminate_clusters
+ */
+uint8_t fat_free_clusters(struct fat_fs_struct* fs, cluster_t cluster_num)
+{
+ if(!fs || cluster_num < 2)
+ return 0;
+
+ offset_t fat_offset = fs->header.fat_offset;
+#if FAT_FAT32_SUPPORT
+ if(fs->partition->type == PARTITION_TYPE_FAT32)
+ {
+ uint32_t fat_entry;
+ while(cluster_num)
+ {
+ if(!fs->partition->device_read(fat_offset + (offset_t) cluster_num * sizeof(fat_entry), (uint8_t*) &fat_entry, sizeof(fat_entry)))
+ return 0;
+
+ /* get next cluster of current cluster before freeing current cluster */
+ uint32_t cluster_num_next = ltoh32(fat_entry);
+
+ if(cluster_num_next == FAT32_CLUSTER_FREE)
+ return 1;
+ if(cluster_num_next == FAT32_CLUSTER_BAD ||
+ (cluster_num_next >= FAT32_CLUSTER_RESERVED_MIN &&
+ cluster_num_next <= FAT32_CLUSTER_RESERVED_MAX
+ )
+ )
+ return 0;
+ if(cluster_num_next >= FAT32_CLUSTER_LAST_MIN && cluster_num_next <= FAT32_CLUSTER_LAST_MAX)
+ cluster_num_next = 0;
+
+ /* We know we will free the cluster, so remember it as
+ * free for the next allocation.
+ */
+ if(!fs->cluster_free)
+ fs->cluster_free = cluster_num;
+
+ /* free cluster */
+ fat_entry = HTOL32(FAT32_CLUSTER_FREE);
+ fs->partition->device_write(fat_offset + (offset_t) cluster_num * sizeof(fat_entry), (uint8_t*) &fat_entry, sizeof(fat_entry));
+
+ /* We continue in any case here, even if freeing the cluster failed.
+ * The cluster is lost, but maybe we can still free up some later ones.
+ */
+
+ cluster_num = cluster_num_next;
+ }
+ }
+ else
+#endif
+ {
+ uint16_t fat_entry;
+ while(cluster_num)
+ {
+ if(!fs->partition->device_read(fat_offset + (offset_t) cluster_num * sizeof(fat_entry), (uint8_t*) &fat_entry, sizeof(fat_entry)))
+ return 0;
+
+ /* get next cluster of current cluster before freeing current cluster */
+ uint16_t cluster_num_next = ltoh16(fat_entry);
+
+ if(cluster_num_next == FAT16_CLUSTER_FREE)
+ return 1;
+ if(cluster_num_next == FAT16_CLUSTER_BAD ||
+ (cluster_num_next >= FAT16_CLUSTER_RESERVED_MIN &&
+ cluster_num_next <= FAT16_CLUSTER_RESERVED_MAX
+ )
+ )
+ return 0;
+ if(cluster_num_next >= FAT16_CLUSTER_LAST_MIN && cluster_num_next <= FAT16_CLUSTER_LAST_MAX)
+ cluster_num_next = 0;
+
+ /* free cluster */
+ fat_entry = HTOL16(FAT16_CLUSTER_FREE);
+ fs->partition->device_write(fat_offset + (offset_t) cluster_num * sizeof(fat_entry), (uint8_t*) &fat_entry, sizeof(fat_entry));
+
+ /* We continue in any case here, even if freeing the cluster failed.
+ * The cluster is lost, but maybe we can still free up some later ones.
+ */
+
+ cluster_num = cluster_num_next;
+ }
+ }
+
+ return 1;
+}
+#endif
+
+#if DOXYGEN || FAT_WRITE_SUPPORT
+/**
+ * \ingroup fat_fs
+ * Frees a part of a cluster chain and correctly terminates the rest.
+ *
+ * Marks the specified cluster as the new end of a cluster chain and
+ * frees all following clusters.
+ *
+ * \param[in] fs The filesystem on which to operate.
+ * \param[in] cluster_num The new end of the cluster chain.
+ * \returns 0 on failure, 1 on success.
+ * \see fat_free_clusters
+ */
+uint8_t fat_terminate_clusters(struct fat_fs_struct* fs, cluster_t cluster_num)
+{
+ if(!fs || cluster_num < 2)
+ return 0;
+
+ /* fetch next cluster before overwriting the cluster entry */
+ cluster_t cluster_num_next = fat_get_next_cluster(fs, cluster_num);
+
+ /* mark cluster as the last one */
+#if FAT_FAT32_SUPPORT
+ if(fs->partition->type == PARTITION_TYPE_FAT32)
+ {
+ uint32_t fat_entry = HTOL32(FAT32_CLUSTER_LAST_MAX);
+ if(!fs->partition->device_write(fs->header.fat_offset + (offset_t) cluster_num * sizeof(fat_entry), (uint8_t*) &fat_entry, sizeof(fat_entry)))
+ return 0;
+ }
+ else
+#endif
+ {
+ uint16_t fat_entry = HTOL16(FAT16_CLUSTER_LAST_MAX);
+ if(!fs->partition->device_write(fs->header.fat_offset + (offset_t) cluster_num * sizeof(fat_entry), (uint8_t*) &fat_entry, sizeof(fat_entry)))
+ return 0;
+ }
+
+ /* free remaining clusters */
+ if(cluster_num_next)
+ return fat_free_clusters(fs, cluster_num_next);
+ else
+ return 1;
+}
+#endif
+
+#if DOXYGEN || FAT_WRITE_SUPPORT
+/**
+ * \ingroup fat_fs
+ * Clears a single cluster.
+ *
+ * The complete cluster is filled with zeros.
+ *
+ * \param[in] fs The filesystem on which to operate.
+ * \param[in] cluster_num The cluster to clear.
+ * \returns 0 on failure, 1 on success.
+ */
+uint8_t fat_clear_cluster(const struct fat_fs_struct* fs, cluster_t cluster_num)
+{
+ if(cluster_num < 2)
+ return 0;
+
+ offset_t cluster_offset = fat_cluster_offset(fs, cluster_num);
+
+ uint8_t zero[16];
+ memset(zero, 0, sizeof(zero));
+ return fs->partition->device_write_interval(cluster_offset,
+ zero,
+ fs->header.cluster_size,
+ fat_clear_cluster_callback,
+ 0
+ );
+}
+#endif
+
+#if DOXYGEN || FAT_WRITE_SUPPORT
+/**
+ * \ingroup fat_fs
+ * Callback function for clearing a cluster.
+ */
+uintptr_t fat_clear_cluster_callback(uint8_t* buffer, offset_t offset, void* p)
+{
+ return 16;
+}
+#endif
+
+/**
+ * \ingroup fat_fs
+ * Calculates the offset of the specified cluster.
+ *
+ * \param[in] fs The filesystem on which to operate.
+ * \param[in] cluster_num The cluster whose offset to calculate.
+ * \returns The cluster offset.
+ */
+offset_t fat_cluster_offset(const struct fat_fs_struct* fs, cluster_t cluster_num)
+{
+ if(!fs || cluster_num < 2)
+ return 0;
+
+ return fs->header.cluster_zero_offset + (offset_t) (cluster_num - 2) * fs->header.cluster_size;
+}
+
+/**
+ * \ingroup fat_file
+ * Retrieves the directory entry of a path.
+ *
+ * The given path may both describe a file or a directory.
+ *
+ * \param[in] fs The FAT filesystem on which to search.
+ * \param[in] path The path of which to read the directory entry.
+ * \param[out] dir_entry The directory entry to fill.
+ * \returns 0 on failure, 1 on success.
+ * \see fat_read_dir
+ */
+uint8_t fat_get_dir_entry_of_path(struct fat_fs_struct* fs, const char* path, struct fat_dir_entry_struct* dir_entry)
+{
+ if(!fs || !path || path[0] == '\0' || !dir_entry)
+ return 0;
+
+ if(path[0] == '/')
+ ++path;
+
+ /* begin with the root directory */
+ memset(dir_entry, 0, sizeof(*dir_entry));
+ dir_entry->attributes = FAT_ATTRIB_DIR;
+
+ while(1)
+ {
+ if(path[0] == '\0')
+ return 1;
+
+ struct fat_dir_struct* dd = fat_open_dir(fs, dir_entry);
+ if(!dd)
+ break;
+
+ /* extract the next hierarchy we will search for */
+ const char* sub_path = strchr(path, '/');
+ uint8_t length_to_sep;
+ if(sub_path)
+ {
+ length_to_sep = sub_path - path;
+ ++sub_path;
+ }
+ else
+ {
+ length_to_sep = strlen(path);
+ sub_path = path + length_to_sep;
+ }
+
+ /* read directory entries */
+ while(fat_read_dir(dd, dir_entry))
+ {
+ /* check if we have found the next hierarchy */
+ if((strlen(dir_entry->long_name) != length_to_sep ||
+ strncmp(path, dir_entry->long_name, length_to_sep) != 0))
+ continue;
+
+ fat_close_dir(dd);
+ dd = 0;
+
+ if(path[length_to_sep] == '\0')
+ /* we iterated through the whole path and have found the file */
+ return 1;
+
+ if(dir_entry->attributes & FAT_ATTRIB_DIR)
+ {
+ /* we found a parent directory of the file we are searching for */
+ path = sub_path;
+ break;
+ }
+
+ /* a parent of the file exists, but not the file itself */
+ return 0;
+ }
+
+ fat_close_dir(dd);
+ }
+
+ return 0;
+}
+
+/**
+ * \ingroup fat_file
+ * Opens a file on a FAT filesystem.
+ *
+ * \param[in] fs The filesystem on which the file to open lies.
+ * \param[in] dir_entry The directory entry of the file to open.
+ * \returns The file handle, or 0 on failure.
+ * \see fat_close_file
+ */
+struct fat_file_struct* fat_open_file(struct fat_fs_struct* fs, const struct fat_dir_entry_struct* dir_entry)
+{
+ if(!fs || !dir_entry || (dir_entry->attributes & FAT_ATTRIB_DIR))
+ return 0;
+
+#if USE_DYNAMIC_MEMORY
+ struct fat_file_struct* fd = malloc(sizeof(*fd));
+ if(!fd)
+ return 0;
+#else
+ struct fat_file_struct* fd = fat_file_handles;
+ uint8_t i;
+ for(i = 0; i < FAT_FILE_COUNT; ++i)
+ {
+ if(!fd->fs)
+ break;
+
+ ++fd;
+ }
+ if(i >= FAT_FILE_COUNT)
+ return 0;
+#endif
+
+ memcpy(&fd->dir_entry, dir_entry, sizeof(*dir_entry));
+ fd->fs = fs;
+ fd->pos = 0;
+ fd->pos_cluster = dir_entry->cluster;
+
+ return fd;
+}
+
+/**
+ * \ingroup fat_file
+ * Closes a file.
+ *
+ * \param[in] fd The file handle of the file to close.
+ * \see fat_open_file
+ */
+void fat_close_file(struct fat_file_struct* fd)
+{
+ if(fd)
+ {
+#if FAT_DELAY_DIRENTRY_UPDATE
+ /* write directory entry */
+ fat_write_dir_entry(fd->fs, &fd->dir_entry);
+#endif
+
+#if USE_DYNAMIC_MEMORY
+ free(fd);
+#else
+ fd->fs = 0;
+#endif
+ }
+}
+
+/**
+ * \ingroup fat_file
+ * Reads data from a file.
+ *
+ * The data requested is read from the current file location.
+ *
+ * \param[in] fd The file handle of the file from which to read.
+ * \param[out] buffer The buffer into which to write.
+ * \param[in] buffer_len The amount of data to read.
+ * \returns The number of bytes read, 0 on end of file, or -1 on failure.
+ * \see fat_write_file
+ */
+intptr_t fat_read_file(struct fat_file_struct* fd, uint8_t* buffer, uintptr_t buffer_len)
+{
+ /* check arguments */
+ if(!fd || !buffer || buffer_len < 1)
+ return -1;
+
+ /* determine number of bytes to read */
+ if(fd->pos + buffer_len > fd->dir_entry.file_size)
+ buffer_len = fd->dir_entry.file_size - fd->pos;
+ if(buffer_len == 0)
+ return 0;
+
+ uint16_t cluster_size = fd->fs->header.cluster_size;
+ cluster_t cluster_num = fd->pos_cluster;
+ uintptr_t buffer_left = buffer_len;
+ uint16_t first_cluster_offset = (uint16_t) (fd->pos & (cluster_size - 1));
+
+ /* find cluster in which to start reading */
+ if(!cluster_num)
+ {
+ cluster_num = fd->dir_entry.cluster;
+
+ if(!cluster_num)
+ {
+ if(!fd->pos)
+ return 0;
+ else
+ return -1;
+ }
+
+ if(fd->pos)
+ {
+ uint32_t pos = fd->pos;
+ while(pos >= cluster_size)
+ {
+ pos -= cluster_size;
+ cluster_num = fat_get_next_cluster(fd->fs, cluster_num);
+ if(!cluster_num)
+ return -1;
+ }
+ }
+ }
+
+ /* read data */
+ do
+ {
+ /* calculate data size to copy from cluster */
+ offset_t cluster_offset = fat_cluster_offset(fd->fs, cluster_num) + first_cluster_offset;
+ uint16_t copy_length = cluster_size - first_cluster_offset;
+ if(copy_length > buffer_left)
+ copy_length = buffer_left;
+
+ /* read data */
+ if(!fd->fs->partition->device_read(cluster_offset, buffer, copy_length))
+ return buffer_len - buffer_left;
+
+ /* calculate new file position */
+ buffer += copy_length;
+ buffer_left -= copy_length;
+ fd->pos += copy_length;
+
+ if(first_cluster_offset + copy_length >= cluster_size)
+ {
+ /* we are on a cluster boundary, so get the next cluster */
+ if((cluster_num = fat_get_next_cluster(fd->fs, cluster_num)))
+ {
+ first_cluster_offset = 0;
+ }
+ else
+ {
+ fd->pos_cluster = 0;
+ return buffer_len - buffer_left;
+ }
+ }
+
+ fd->pos_cluster = cluster_num;
+
+ } while(buffer_left > 0); /* check if we are done */
+
+ return buffer_len;
+}
+
+#if DOXYGEN || FAT_WRITE_SUPPORT
+/**
+ * \ingroup fat_file
+ * Writes data to a file.
+ *
+ * The data is written to the current file location.
+ *
+ * \param[in] fd The file handle of the file to which to write.
+ * \param[in] buffer The buffer from which to read the data to be written.
+ * \param[in] buffer_len The amount of data to write.
+ * \returns The number of bytes written (0 or something less than \c buffer_len on disk full) or -1 on failure.
+ * \see fat_read_file
+ */
+intptr_t fat_write_file(struct fat_file_struct* fd, const uint8_t* buffer, uintptr_t buffer_len)
+{
+ /* check arguments */
+ if(!fd || !buffer || buffer_len < 1)
+ return -1;
+ if(fd->pos > fd->dir_entry.file_size)
+ return -1;
+
+ uint16_t cluster_size = fd->fs->header.cluster_size;
+ cluster_t cluster_num = fd->pos_cluster;
+ uintptr_t buffer_left = buffer_len;
+ uint16_t first_cluster_offset = (uint16_t) (fd->pos & (cluster_size - 1));
+
+ /* find cluster in which to start writing */
+ if(!cluster_num)
+ {
+ cluster_num = fd->dir_entry.cluster;
+
+ if(!cluster_num)
+ {
+ if(!fd->pos)
+ {
+ /* empty file */
+ fd->dir_entry.cluster = cluster_num = fat_append_clusters(fd->fs, 0, 1);
+ if(!cluster_num)
+ return 0;
+ }
+ else
+ {
+ return -1;
+ }
+ }
+
+ if(fd->pos)
+ {
+ uint32_t pos = fd->pos;
+ cluster_t cluster_num_next;
+ while(pos >= cluster_size)
+ {
+ pos -= cluster_size;
+ cluster_num_next = fat_get_next_cluster(fd->fs, cluster_num);
+ if(!cluster_num_next)
+ {
+ if(pos != 0)
+ return -1; /* current file position points beyond end of file */
+
+ /* the file exactly ends on a cluster boundary, and we append to it */
+ cluster_num_next = fat_append_clusters(fd->fs, cluster_num, 1);
+ if(!cluster_num_next)
+ return 0;
+ }
+
+ cluster_num = cluster_num_next;
+ }
+ }
+ }
+
+ /* write data */
+ do
+ {
+ /* calculate data size to write to cluster */
+ offset_t cluster_offset = fat_cluster_offset(fd->fs, cluster_num) + first_cluster_offset;
+ uint16_t write_length = cluster_size - first_cluster_offset;
+ if(write_length > buffer_left)
+ write_length = buffer_left;
+
+ /* write data which fits into the current cluster */
+ if(!fd->fs->partition->device_write(cluster_offset, buffer, write_length))
+ break;
+
+ /* calculate new file position */
+ buffer += write_length;
+ buffer_left -= write_length;
+ fd->pos += write_length;
+
+ if(first_cluster_offset + write_length >= cluster_size)
+ {
+ /* we are on a cluster boundary, so get the next cluster */
+ cluster_t cluster_num_next = fat_get_next_cluster(fd->fs, cluster_num);
+ if(!cluster_num_next && buffer_left > 0)
+ /* we reached the last cluster, append a new one */
+ cluster_num_next = fat_append_clusters(fd->fs, cluster_num, 1);
+ if(!cluster_num_next)
+ {
+ fd->pos_cluster = 0;
+ break;
+ }
+
+ cluster_num = cluster_num_next;
+ first_cluster_offset = 0;
+ }
+
+ fd->pos_cluster = cluster_num;
+
+ } while(buffer_left > 0); /* check if we are done */
+
+ /* update directory entry */
+ if(fd->pos > fd->dir_entry.file_size)
+ {
+#if !FAT_DELAY_DIRENTRY_UPDATE
+ uint32_t size_old = fd->dir_entry.file_size;
+#endif
+
+ /* update file size */
+ fd->dir_entry.file_size = fd->pos;
+
+#if !FAT_DELAY_DIRENTRY_UPDATE
+ /* write directory entry */
+ if(!fat_write_dir_entry(fd->fs, &fd->dir_entry))
+ {
+ /* We do not return an error here since we actually wrote
+ * some data to disk. So we calculate the amount of data
+ * we wrote to disk and which lies within the old file size.
+ */
+ buffer_left = fd->pos - size_old;
+ fd->pos = size_old;
+ }
+#endif
+ }
+
+ return buffer_len - buffer_left;
+}
+#endif
+
+/**
+ * \ingroup fat_file
+ * Repositions the read/write file offset.
+ *
+ * Changes the file offset where the next call to fat_read_file()
+ * or fat_write_file() starts reading/writing.
+ *
+ * If the new offset is beyond the end of the file, fat_resize_file()
+ * is implicitly called, i.e. the file is expanded.
+ *
+ * The new offset can be given in different ways determined by
+ * the \c whence parameter:
+ * - \b FAT_SEEK_SET: \c *offset is relative to the beginning of the file.
+ * - \b FAT_SEEK_CUR: \c *offset is relative to the current file position.
+ * - \b FAT_SEEK_END: \c *offset is relative to the end of the file.
+ *
+ * The resulting absolute offset is written to the location the \c offset
+ * parameter points to.
+ *
+ * Calling this function can also be used to retrieve the current file position:
+ \code
+ int32_t file_pos = 0;
+ if(!fat_seek_file(fd, &file_pos, FAT_SEEK_CUR))
+ {
+ // error
+ }
+ // file_pos now contains the absolute file position
+ \endcode
+ *
+ * \param[in] fd The file decriptor of the file on which to seek.
+ * \param[in,out] offset A pointer to the new offset, as affected by the \c whence
+ * parameter. The function writes the new absolute offset
+ * to this location before it returns.
+ * \param[in] whence Affects the way \c offset is interpreted, see above.
+ * \returns 0 on failure, 1 on success.
+ */
+uint8_t fat_seek_file(struct fat_file_struct* fd, int32_t* offset, uint8_t whence)
+{
+ if(!fd || !offset)
+ return 0;
+
+ uint32_t new_pos = fd->pos;
+ switch(whence)
+ {
+ case FAT_SEEK_SET:
+ new_pos = *offset;
+ break;
+ case FAT_SEEK_CUR:
+ new_pos += *offset;
+ break;
+ case FAT_SEEK_END:
+ new_pos = fd->dir_entry.file_size + *offset;
+ break;
+ default:
+ return 0;
+ }
+
+ if(new_pos > fd->dir_entry.file_size
+#if FAT_WRITE_SUPPORT
+ && !fat_resize_file(fd, new_pos)
+#endif
+ )
+ return 0;
+
+ fd->pos = new_pos;
+ fd->pos_cluster = 0;
+
+ *offset = (int32_t) new_pos;
+ return 1;
+}
+
+#if DOXYGEN || FAT_WRITE_SUPPORT
+/**
+ * \ingroup fat_file
+ * Resizes a file to have a specific size.
+ *
+ * Enlarges or shrinks the file pointed to by the file descriptor to have
+ * exactly the specified size.
+ *
+ * If the file is truncated, all bytes having an equal or larger offset
+ * than the given size are lost. If the file is expanded, the additional
+ * bytes are allocated.
+ *
+ * \note Please be aware that this function just allocates or deallocates disk
+ * space, it does not explicitely clear it. To avoid data leakage, this
+ * must be done manually.
+ *
+ * \param[in] fd The file decriptor of the file which to resize.
+ * \param[in] size The new size of the file.
+ * \returns 0 on failure, 1 on success.
+ */
+uint8_t fat_resize_file(struct fat_file_struct* fd, uint32_t size)
+{
+ if(!fd)
+ return 0;
+
+ cluster_t cluster_num = fd->dir_entry.cluster;
+ uint16_t cluster_size = fd->fs->header.cluster_size;
+ uint32_t size_new = size;
+
+ do
+ {
+ if(cluster_num == 0 && size_new == 0)
+ /* the file stays empty */
+ break;
+
+ /* seek to the next cluster as long as we need the space */
+ while(size_new > cluster_size)
+ {
+ /* get next cluster of file */
+ cluster_t cluster_num_next = fat_get_next_cluster(fd->fs, cluster_num);
+ if(cluster_num_next)
+ {
+ cluster_num = cluster_num_next;
+ size_new -= cluster_size;
+ }
+ else
+ {
+ break;
+ }
+ }
+
+ if(size_new > cluster_size || cluster_num == 0)
+ {
+ /* Allocate new cluster chain and append
+ * it to the existing one, if available.
+ */
+ cluster_t cluster_count = (size_new + cluster_size - 1) / cluster_size;
+ cluster_t cluster_new_chain = fat_append_clusters(fd->fs, cluster_num, cluster_count);
+ if(!cluster_new_chain)
+ return 0;
+
+ if(!cluster_num)
+ {
+ cluster_num = cluster_new_chain;
+ fd->dir_entry.cluster = cluster_num;
+ }
+ }
+
+ /* write new directory entry */
+ fd->dir_entry.file_size = size;
+ if(size == 0)
+ fd->dir_entry.cluster = 0;
+ if(!fat_write_dir_entry(fd->fs, &fd->dir_entry))
+ return 0;
+
+ if(size == 0)
+ {
+ /* free all clusters of file */
+ fat_free_clusters(fd->fs, cluster_num);
+ }
+ else if(size_new <= cluster_size)
+ {
+ /* free all clusters no longer needed */
+ fat_terminate_clusters(fd->fs, cluster_num);
+ }
+
+ } while(0);
+
+ /* correct file position */
+ if(size < fd->pos)
+ {
+ fd->pos = size;
+ fd->pos_cluster = 0;
+ }
+
+ return 1;
+}
+#endif
+
+/**
+ * \ingroup fat_dir
+ * Opens a directory.
+ *
+ * \param[in] fs The filesystem on which the directory to open resides.
+ * \param[in] dir_entry The directory entry which stands for the directory to open.
+ * \returns An opaque directory descriptor on success, 0 on failure.
+ * \see fat_close_dir
+ */
+struct fat_dir_struct* fat_open_dir(struct fat_fs_struct* fs, const struct fat_dir_entry_struct* dir_entry)
+{
+ if(!fs || !dir_entry || !(dir_entry->attributes & FAT_ATTRIB_DIR))
+ return 0;
+
+#if USE_DYNAMIC_MEMORY
+ struct fat_dir_struct* dd = malloc(sizeof(*dd));
+ if(!dd)
+ return 0;
+#else
+ struct fat_dir_struct* dd = fat_dir_handles;
+ uint8_t i;
+ for(i = 0; i < FAT_DIR_COUNT; ++i)
+ {
+ if(!dd->fs)
+ break;
+
+ ++dd;
+ }
+ if(i >= FAT_DIR_COUNT)
+ return 0;
+#endif
+
+ memcpy(&dd->dir_entry, dir_entry, sizeof(*dir_entry));
+ dd->fs = fs;
+ dd->entry_cluster = dir_entry->cluster;
+ dd->entry_offset = 0;
+
+ return dd;
+}
+
+/**
+ * \ingroup fat_dir
+ * Closes a directory descriptor.
+ *
+ * This function destroys a directory descriptor which was
+ * previously obtained by calling fat_open_dir(). When this
+ * function returns, the given descriptor will be invalid.
+ *
+ * \param[in] dd The directory descriptor to close.
+ * \see fat_open_dir
+ */
+void fat_close_dir(struct fat_dir_struct* dd)
+{
+ if(dd)
+#if USE_DYNAMIC_MEMORY
+ free(dd);
+#else
+ dd->fs = 0;
+#endif
+}
+
+/**
+ * \ingroup fat_dir
+ * Reads the next directory entry contained within a parent directory.
+ *
+ * \param[in] dd The descriptor of the parent directory from which to read the entry.
+ * \param[out] dir_entry Pointer to a buffer into which to write the directory entry information.
+ * \returns 0 on failure, 1 on success.
+ * \see fat_reset_dir
+ */
+uint8_t fat_read_dir(struct fat_dir_struct* dd, struct fat_dir_entry_struct* dir_entry)
+{
+ if(!dd || !dir_entry)
+ return 0;
+
+ /* get current position of directory handle */
+ struct fat_fs_struct* fs = dd->fs;
+ const struct fat_header_struct* header = &fs->header;
+ uint16_t cluster_size = header->cluster_size;
+ cluster_t cluster_num = dd->entry_cluster;
+ uint16_t cluster_offset = dd->entry_offset;
+ struct fat_read_dir_callback_arg arg;
+
+ if(cluster_offset >= cluster_size)
+ {
+ /* The latest call hit the border of the last cluster in
+ * the chain, but it still returned a directory entry.
+ * So we now reset the handle and signal the caller the
+ * end of the listing.
+ */
+ fat_reset_dir(dd);
+ return 0;
+ }
+
+ /* reset callback arguments */
+ memset(&arg, 0, sizeof(arg));
+ memset(dir_entry, 0, sizeof(*dir_entry));
+ arg.dir_entry = dir_entry;
+
+ /* check if we read from the root directory */
+ if(cluster_num == 0)
+ {
+#if FAT_FAT32_SUPPORT
+ if(fs->partition->type == PARTITION_TYPE_FAT32)
+ cluster_num = header->root_dir_cluster;
+ else
+#endif
+ cluster_size = header->cluster_zero_offset - header->root_dir_offset;
+ }
+
+ /* read entries */
+ uint8_t buffer[32];
+ while(!arg.finished)
+ {
+ /* read directory entries up to the cluster border */
+ uint16_t cluster_left = cluster_size - cluster_offset;
+ offset_t pos = cluster_offset;
+ if(cluster_num == 0)
+ pos += header->root_dir_offset;
+ else
+ pos += fat_cluster_offset(fs, cluster_num);
+
+ arg.bytes_read = 0;
+ if(!fs->partition->device_read_interval(pos,
+ buffer,
+ sizeof(buffer),
+ cluster_left,
+ fat_dir_entry_read_callback,
+ &arg)
+ )
+ return 0;
+
+ cluster_offset += arg.bytes_read;
+
+ if(cluster_offset >= cluster_size)
+ {
+ /* we reached the cluster border and switch to the next cluster */
+
+ /* get number of next cluster */
+ if((cluster_num = fat_get_next_cluster(fs, cluster_num)) != 0)
+ {
+ cluster_offset = 0;
+ continue;
+ }
+
+ /* we are at the end of the cluster chain */
+ if(!arg.finished)
+ {
+ /* directory entry not found, reset directory handle */
+ fat_reset_dir(dd);
+ return 0;
+ }
+ else
+ {
+ /* The current execution of the function has been successful,
+ * so we can not signal an end of the directory listing to
+ * the caller, but must wait for the next call. So we keep an
+ * invalid cluster offset to mark this directory handle's
+ * traversal as finished.
+ */
+ }
+
+ break;
+ }
+ }
+
+ dd->entry_cluster = cluster_num;
+ dd->entry_offset = cluster_offset;
+
+ return arg.finished;
+}
+
+/**
+ * \ingroup fat_dir
+ * Resets a directory handle.
+ *
+ * Resets the directory handle such that reading restarts
+ * with the first directory entry.
+ *
+ * \param[in] dd The directory handle to reset.
+ * \returns 0 on failure, 1 on success.
+ * \see fat_read_dir
+ */
+uint8_t fat_reset_dir(struct fat_dir_struct* dd)
+{
+ if(!dd)
+ return 0;
+
+ dd->entry_cluster = dd->dir_entry.cluster;
+ dd->entry_offset = 0;
+ return 1;
+}
+
+/**
+ * \ingroup fat_fs
+ * Callback function for reading a directory entry.
+ *
+ * Interprets a raw directory entry and puts the contained
+ * information into a fat_dir_entry_struct structure.
+ *
+ * For a single file there may exist multiple directory
+ * entries. All except the last one are lfn entries, which
+ * contain parts of the long filename. The last directory
+ * entry is a traditional 8.3 style one. It contains all
+ * other information like size, cluster, date and time.
+ *
+ * \param[in] buffer A pointer to 32 bytes of raw data.
+ * \param[in] offset The absolute offset of the raw data.
+ * \param[in,out] p An argument structure controlling operation.
+ * \returns 0 on failure or completion, 1 if reading has
+ * to be continued
+ */
+uint8_t fat_dir_entry_read_callback(uint8_t* buffer, offset_t offset, void* p)
+{
+ struct fat_read_dir_callback_arg* arg = p;
+ struct fat_dir_entry_struct* dir_entry = arg->dir_entry;
+
+ arg->bytes_read += 32;
+
+ /* skip deleted or empty entries */
+ if(buffer[0] == FAT_DIRENTRY_DELETED || !buffer[0])
+ {
+#if FAT_LFN_SUPPORT
+ arg->checksum = 0;
+#endif
+ return 1;
+ }
+
+#if !FAT_LFN_SUPPORT
+ /* skip lfn entries */
+ if(buffer[11] == 0x0f)
+ return 1;
+#endif
+
+ char* long_name = dir_entry->long_name;
+#if FAT_LFN_SUPPORT
+ if(buffer[11] == 0x0f)
+ {
+ /* checksum validation */
+ if(arg->checksum == 0 || arg->checksum != buffer[13])
+ {
+ /* reset directory entry */
+ memset(dir_entry, 0, sizeof(*dir_entry));
+
+ arg->checksum = buffer[13];
+ dir_entry->entry_offset = offset;
+ }
+
+ /* lfn supports unicode, but we do not, for now.
+ * So we assume pure ascii and read only every
+ * second byte.
+ */
+ uint16_t char_offset = ((buffer[0] & 0x3f) - 1) * 13;
+ const uint8_t char_mapping[] = { 1, 3, 5, 7, 9, 14, 16, 18, 20, 22, 24, 28, 30 };
+ for(uint8_t i = 0; i <= 12 && char_offset + i < sizeof(dir_entry->long_name) - 1; ++i)
+ long_name[char_offset + i] = buffer[char_mapping[i]];
+
+ return 1;
+ }
+ else
+#endif
+ {
+#if FAT_LFN_SUPPORT
+ /* if we do not have a long name or the previous lfn does not match, take the 8.3 name */
+ if(long_name[0] == '\0' || arg->checksum != fat_calc_83_checksum(buffer))
+#endif
+ {
+ /* reset directory entry */
+ memset(dir_entry, 0, sizeof(*dir_entry));
+ dir_entry->entry_offset = offset;
+
+ uint8_t i;
+ for(i = 0; i < 8; ++i)
+ {
+ if(buffer[i] == ' ')
+ break;
+ long_name[i] = buffer[i];
+
+ /* Windows NT and later versions do not store lfn entries
+ * for 8.3 names which have a lowercase basename, extension
+ * or both when everything else is uppercase. They use two
+ * extra bits to signal a lowercase basename or extension.
+ */
+ if((buffer[12] & 0x08) && buffer[i] >= 'A' && buffer[i] <= 'Z')
+ long_name[i] += 'a' - 'A';
+ }
+ if(long_name[0] == 0x05)
+ long_name[0] = (char) FAT_DIRENTRY_DELETED;
+
+ if(buffer[8] != ' ')
+ {
+ long_name[i++] = '.';
+
+ uint8_t j = 8;
+ for(; j < 11; ++j)
+ {
+ if(buffer[j] == ' ')
+ break;
+ long_name[i] = buffer[j];
+
+ /* See above for the lowercase 8.3 name handling of
+ * Windows NT and later.
+ */
+ if((buffer[12] & 0x10) && buffer[j] >= 'A' && buffer[j] <= 'Z')
+ long_name[i] += 'a' - 'A';
+
+ ++i;
+ }
+ }
+
+ long_name[i] = '\0';
+ }
+
+ /* extract properties of file and store them within the structure */
+ dir_entry->attributes = buffer[11];
+ dir_entry->cluster = read16(&buffer[26]);
+#if FAT_FAT32_SUPPORT
+ dir_entry->cluster |= ((cluster_t) read16(&buffer[20])) << 16;
+#endif
+ dir_entry->file_size = read32(&buffer[28]);
+
+#if FAT_DATETIME_SUPPORT
+ dir_entry->modification_time = read16(&buffer[22]);
+ dir_entry->modification_date = read16(&buffer[24]);
+#endif
+
+ arg->finished = 1;
+ return 0;
+ }
+}
+
+#if DOXYGEN || FAT_LFN_SUPPORT
+/**
+ * \ingroup fat_fs
+ * Calculates the checksum for 8.3 names used within the
+ * corresponding lfn directory entries.
+ *
+ * \param[in] file_name_83 The 11-byte file name buffer.
+ * \returns The checksum of the given file name.
+ */
+uint8_t fat_calc_83_checksum(const uint8_t* file_name_83)
+{
+ uint8_t checksum = file_name_83[0];
+ for(uint8_t i = 1; i < 11; ++i)
+ checksum = ((checksum >> 1) | (checksum << 7)) + file_name_83[i];
+
+ return checksum;
+}
+#endif
+
+#if DOXYGEN || FAT_WRITE_SUPPORT
+/**
+ * \ingroup fat_fs
+ * Searches for space where to store a directory entry.
+ *
+ * \param[in] fs The filesystem on which to operate.
+ * \param[in] parent The directory in which to search.
+ * \param[in] dir_entry The directory entry for which to search space.
+ * \returns 0 on failure, a device offset on success.
+ */
+offset_t fat_find_offset_for_dir_entry(struct fat_fs_struct* fs, const struct fat_dir_struct* parent, const struct fat_dir_entry_struct* dir_entry)
+{
+ if(!fs || !dir_entry)
+ return 0;
+
+ /* search for a place where to write the directory entry to disk */
+#if FAT_LFN_SUPPORT
+ uint8_t free_dir_entries_needed = (strlen(dir_entry->long_name) + 12) / 13 + 1;
+ uint8_t free_dir_entries_found = 0;
+#endif
+ cluster_t cluster_num = parent->dir_entry.cluster;
+ offset_t dir_entry_offset = 0;
+ offset_t offset = 0;
+ offset_t offset_to = 0;
+#if FAT_FAT32_SUPPORT
+ uint8_t is_fat32 = (fs->partition->type == PARTITION_TYPE_FAT32);
+#endif
+
+ if(cluster_num == 0)
+ {
+#if FAT_FAT32_SUPPORT
+ if(is_fat32)
+ {
+ cluster_num = fs->header.root_dir_cluster;
+ }
+ else
+#endif
+ {
+ /* we read/write from the root directory entry */
+ offset = fs->header.root_dir_offset;
+ offset_to = fs->header.cluster_zero_offset;
+ dir_entry_offset = offset;
+ }
+ }
+
+ while(1)
+ {
+ if(offset == offset_to)
+ {
+ if(cluster_num == 0)
+ /* We iterated through the whole root directory and
+ * could not find enough space for the directory entry.
+ */
+ return 0;
+
+ if(offset)
+ {
+ /* We reached a cluster boundary and have to
+ * switch to the next cluster.
+ */
+
+ cluster_t cluster_next = fat_get_next_cluster(fs, cluster_num);
+ if(!cluster_next)
+ {
+ cluster_next = fat_append_clusters(fs, cluster_num, 1);
+ if(!cluster_next)
+ return 0;
+
+ /* we appended a new cluster and know it is free */
+ dir_entry_offset = fs->header.cluster_zero_offset +
+ (offset_t) (cluster_next - 2) * fs->header.cluster_size;
+
+ /* clear cluster to avoid garbage directory entries */
+ fat_clear_cluster(fs, cluster_next);
+
+ break;
+ }
+ cluster_num = cluster_next;
+ }
+
+ offset = fat_cluster_offset(fs, cluster_num);
+ offset_to = offset + fs->header.cluster_size;
+ dir_entry_offset = offset;
+#if FAT_LFN_SUPPORT
+ free_dir_entries_found = 0;
+#endif
+ }
+
+ /* read next lfn or 8.3 entry */
+ uint8_t first_char;
+ if(!fs->partition->device_read(offset, &first_char, sizeof(first_char)))
+ return 0;
+
+ /* check if we found a free directory entry */
+ if(first_char == FAT_DIRENTRY_DELETED || !first_char)
+ {
+ /* check if we have the needed number of available entries */
+#if FAT_LFN_SUPPORT
+ ++free_dir_entries_found;
+ if(free_dir_entries_found >= free_dir_entries_needed)
+#endif
+ break;
+
+ offset += 32;
+ }
+ else
+ {
+ offset += 32;
+ dir_entry_offset = offset;
+#if FAT_LFN_SUPPORT
+ free_dir_entries_found = 0;
+#endif
+ }
+ }
+
+ return dir_entry_offset;
+}
+#endif
+
+#if DOXYGEN || FAT_WRITE_SUPPORT
+/**
+ * \ingroup fat_fs
+ * Writes a directory entry to disk.
+ *
+ * \note The file name is not checked for invalid characters.
+ *
+ * \note The generation of the short 8.3 file name is quite
+ * simple. The first eight characters are used for the filename.
+ * The extension, if any, is made up of the first three characters
+ * following the last dot within the long filename. If the
+ * filename (without the extension) is longer than eight characters,
+ * the lower byte of the cluster number replaces the last two
+ * characters to avoid name clashes. In any other case, it is your
+ * responsibility to avoid name clashes.
+ *
+ * \param[in] fs The filesystem on which to operate.
+ * \param[in] dir_entry The directory entry to write.
+ * \returns 0 on failure, 1 on success.
+ */
+uint8_t fat_write_dir_entry(const struct fat_fs_struct* fs, struct fat_dir_entry_struct* dir_entry)
+{
+ if(!fs || !dir_entry)
+ return 0;
+
+#if FAT_DATETIME_SUPPORT
+ {
+ uint16_t year;
+ uint8_t month;
+ uint8_t day;
+ uint8_t hour;
+ uint8_t min;
+ uint8_t sec;
+
+ fat_get_datetime(&year, &month, &day, &hour, &min, &sec);
+ fat_set_file_modification_date(dir_entry, year, month, day);
+ fat_set_file_modification_time(dir_entry, hour, min, sec);
+ }
+#endif
+
+ device_write_t device_write = fs->partition->device_write;
+ offset_t offset = dir_entry->entry_offset;
+ const char* name = dir_entry->long_name;
+ uint8_t name_len = strlen(name);
+#if FAT_LFN_SUPPORT
+ uint8_t lfn_entry_count = (name_len + 12) / 13;
+#endif
+ uint8_t buffer[32];
+
+ /* write 8.3 entry */
+
+ /* generate 8.3 file name */
+ memset(&buffer[0], ' ', 11);
+ char* name_ext = strrchr(name, '.');
+ if(name_ext && *++name_ext)
+ {
+ uint8_t name_ext_len = strlen(name_ext);
+ name_len -= name_ext_len + 1;
+
+ if(name_ext_len > 3)
+#if FAT_LFN_SUPPORT
+ name_ext_len = 3;
+#else
+ return 0;
+#endif
+
+ memcpy(&buffer[8], name_ext, name_ext_len);
+ }
+
+ if(name_len <= 8)
+ {
+ memcpy(buffer, name, name_len);
+
+#if FAT_LFN_SUPPORT
+ /* For now, we create lfn entries for all files,
+ * except the "." and ".." directory references.
+ * This is to avoid difficulties with capitalization,
+ * as 8.3 filenames allow uppercase letters only.
+ *
+ * Theoretically it would be possible to leave
+ * the 8.3 entry alone if the basename and the
+ * extension have no mixed capitalization.
+ */
+ if(name[0] == '.' &&
+ ((name[1] == '.' && name[2] == '\0') ||
+ name[1] == '\0')
+ )
+ lfn_entry_count = 0;
+#endif
+ }
+ else
+ {
+#if FAT_LFN_SUPPORT
+ memcpy(buffer, name, 8);
+
+ /* Minimize 8.3 name clashes by appending
+ * the lower byte of the cluster number.
+ */
+ uint8_t num = dir_entry->cluster & 0xff;
+
+ buffer[6] = (num < 0xa0) ? ('0' + (num >> 4)) : ('a' + (num >> 4));
+ num &= 0x0f;
+ buffer[7] = (num < 0x0a) ? ('0' + num) : ('a' + num);
+#else
+ return 0;
+#endif
+ }
+ if(buffer[0] == FAT_DIRENTRY_DELETED)
+ buffer[0] = 0x05;
+
+ /* fill directory entry buffer */
+ memset(&buffer[11], 0, sizeof(buffer) - 11);
+ buffer[0x0b] = dir_entry->attributes;
+#if FAT_DATETIME_SUPPORT
+ write16(&buffer[0x16], dir_entry->modification_time);
+ write16(&buffer[0x18], dir_entry->modification_date);
+#endif
+#if FAT_FAT32_SUPPORT
+ write16(&buffer[0x14], (uint16_t) (dir_entry->cluster >> 16));
+#endif
+ write16(&buffer[0x1a], dir_entry->cluster);
+ write32(&buffer[0x1c], dir_entry->file_size);
+
+ /* write to disk */
+#if FAT_LFN_SUPPORT
+ if(!device_write(offset + (uint16_t) lfn_entry_count * 32, buffer, sizeof(buffer)))
+#else
+ if(!device_write(offset, buffer, sizeof(buffer)))
+#endif
+ return 0;
+
+#if FAT_LFN_SUPPORT
+ /* calculate checksum of 8.3 name */
+ uint8_t checksum = fat_calc_83_checksum(buffer);
+
+ /* write lfn entries */
+ for(uint8_t lfn_entry = lfn_entry_count; lfn_entry > 0; --lfn_entry)
+ {
+ memset(buffer, 0xff, sizeof(buffer));
+
+ /* set file name */
+ const char* long_name_curr = name + (lfn_entry - 1) * 13;
+ uint8_t i = 1;
+ while(i < 0x1f)
+ {
+ buffer[i++] = *long_name_curr;
+ buffer[i++] = 0;
+
+ switch(i)
+ {
+ case 0x0b:
+ i = 0x0e;
+ break;
+ case 0x1a:
+ i = 0x1c;
+ break;
+ }
+
+ if(!*long_name_curr++)
+ break;
+ }
+
+ /* set index of lfn entry */
+ buffer[0x00] = lfn_entry;
+ if(lfn_entry == lfn_entry_count)
+ buffer[0x00] |= FAT_DIRENTRY_LFNLAST;
+
+ /* mark as lfn entry */
+ buffer[0x0b] = 0x0f;
+
+ /* set 8.3 checksum */
+ buffer[0x0d] = checksum;
+
+ /* clear reserved bytes */
+ buffer[0x0c] = 0;
+ buffer[0x1a] = 0;
+ buffer[0x1b] = 0;
+
+ /* write entry */
+ device_write(offset, buffer, sizeof(buffer));
+
+ offset += sizeof(buffer);
+ }
+#endif
+
+ return 1;
+}
+#endif
+
+#if DOXYGEN || FAT_WRITE_SUPPORT
+/**
+ * \ingroup fat_file
+ * Creates a file.
+ *
+ * Creates a file and obtains the directory entry of the
+ * new file. If the file to create already exists, the
+ * directory entry of the existing file will be returned
+ * within the dir_entry parameter.
+ *
+ * \note The file name is not checked for invalid characters.
+ *
+ * \note The generation of the short 8.3 file name is quite
+ * simple. The first eight characters are used for the filename.
+ * The extension, if any, is made up of the first three characters
+ * following the last dot within the long filename. If the
+ * filename (without the extension) is longer than eight characters,
+ * the lower byte of the cluster number replaces the last two
+ * characters to avoid name clashes. In any other case, it is your
+ * responsibility to avoid name clashes.
+ *
+ * \param[in] parent The handle of the directory in which to create the file.
+ * \param[in] file The name of the file to create.
+ * \param[out] dir_entry The directory entry to fill for the new (or existing) file.
+ * \returns 0 on failure, 1 on success, 2 if the file already existed.
+ * \see fat_delete_file
+ */
+uint8_t fat_create_file(struct fat_dir_struct* parent, const char* file, struct fat_dir_entry_struct* dir_entry)
+{
+ if(!parent || !file || !file[0] || !dir_entry)
+ return 0;
+
+ /* check if the file already exists */
+ while(1)
+ {
+ if(!fat_read_dir(parent, dir_entry))
+ break;
+
+ if(strcmp(file, dir_entry->long_name) == 0)
+ {
+ fat_reset_dir(parent);
+ return 2;
+ }
+ }
+
+ struct fat_fs_struct* fs = parent->fs;
+
+ /* prepare directory entry with values already known */
+ memset(dir_entry, 0, sizeof(*dir_entry));
+ strncpy(dir_entry->long_name, file, sizeof(dir_entry->long_name) - 1);
+
+ /* find place where to store directory entry */
+ if(!(dir_entry->entry_offset = fat_find_offset_for_dir_entry(fs, parent, dir_entry)))
+ return 0;
+
+ /* write directory entry to disk */
+ if(!fat_write_dir_entry(fs, dir_entry))
+ return 0;
+
+ return 1;
+}
+#endif
+
+#if DOXYGEN || FAT_WRITE_SUPPORT
+/**
+ * \ingroup fat_file
+ * Deletes a file or directory.
+ *
+ * If a directory is deleted without first deleting its
+ * subdirectories and files, disk space occupied by these
+ * files will get wasted as there is no chance to release
+ * it and mark it as free.
+ *
+ * \param[in] fs The filesystem on which to operate.
+ * \param[in] dir_entry The directory entry of the file to delete.
+ * \returns 0 on failure, 1 on success.
+ * \see fat_create_file
+ */
+uint8_t fat_delete_file(struct fat_fs_struct* fs, struct fat_dir_entry_struct* dir_entry)
+{
+ if(!fs || !dir_entry)
+ return 0;
+
+ /* get offset of the file's directory entry */
+ offset_t dir_entry_offset = dir_entry->entry_offset;
+ if(!dir_entry_offset)
+ return 0;
+
+#if FAT_LFN_SUPPORT
+ uint8_t buffer[12];
+ while(1)
+ {
+ /* read directory entry */
+ if(!fs->partition->device_read(dir_entry_offset, buffer, sizeof(buffer)))
+ return 0;
+
+ /* mark the directory entry as deleted */
+ buffer[0] = FAT_DIRENTRY_DELETED;
+
+ /* write back entry */
+ if(!fs->partition->device_write(dir_entry_offset, buffer, sizeof(buffer)))
+ return 0;
+
+ /* check if we deleted the whole entry */
+ if(buffer[11] != 0x0f)
+ break;
+
+ dir_entry_offset += 32;
+ }
+#else
+ /* mark the directory entry as deleted */
+ uint8_t first_char = FAT_DIRENTRY_DELETED;
+ if(!fs->partition->device_write(dir_entry_offset, &first_char, 1))
+ return 0;
+#endif
+
+ /* We deleted the directory entry. The next thing to do is
+ * marking all occupied clusters as free.
+ */
+ return (dir_entry->cluster == 0 || fat_free_clusters(fs, dir_entry->cluster));
+}
+#endif
+
+#if DOXYGEN || FAT_WRITE_SUPPORT
+/**
+ * \ingroup fat_file
+ * Moves or renames a file.
+ *
+ * Changes a file's name, optionally moving it into another
+ * directory as well. Before calling this function, the
+ * target file name must not exist. Moving a file to a
+ * different filesystem (i.e. \a parent_new doesn't lie on
+ * \a fs) is not supported.
+ *
+ * After successfully renaming (and moving) the file, the
+ * given directory entry is updated such that it points to
+ * the file's new location.
+ *
+ * \note The notes which apply to fat_create_file() also
+ * apply to this function.
+ *
+ * \param[in] fs The filesystem on which to operate.
+ * \param[in,out] dir_entry The directory entry of the file to move.
+ * \param[in] parent_new The handle of the new parent directory of the file.
+ * \param[in] file_new The file's new name.
+ * \returns 0 on failure, 1 on success.
+ * \see fat_create_file, fat_delete_file, fat_move_dir
+ */
+uint8_t fat_move_file(struct fat_fs_struct* fs, struct fat_dir_entry_struct* dir_entry, struct fat_dir_struct* parent_new, const char* file_new)
+{
+ if(!fs || !dir_entry || !parent_new || (file_new && !file_new[0]))
+ return 0;
+ if(fs != parent_new->fs)
+ return 0;
+
+ /* use existing file name if none has been specified */
+ if(!file_new)
+ file_new = dir_entry->long_name;
+
+ /* create file with new file name */
+ struct fat_dir_entry_struct dir_entry_new;
+ if(!fat_create_file(parent_new, file_new, &dir_entry_new))
+ return 0;
+
+ /* copy members of directory entry which do not change with rename */
+ dir_entry_new.attributes = dir_entry->attributes;
+#if FAT_DATETIME_SUPPORT
+ dir_entry_new.modification_time = dir_entry->modification_time;
+ dir_entry_new.modification_date = dir_entry->modification_date;
+#endif
+ dir_entry_new.cluster = dir_entry->cluster;
+ dir_entry_new.file_size = dir_entry->file_size;
+
+ /* make the new file name point to the old file's content */
+ if(!fat_write_dir_entry(fs, &dir_entry_new))
+ {
+ fat_delete_file(fs, &dir_entry_new);
+ return 0;
+ }
+
+ /* delete the old file, but not its clusters, which have already been remapped above */
+ dir_entry->cluster = 0;
+ if(!fat_delete_file(fs, dir_entry))
+ return 0;
+
+ *dir_entry = dir_entry_new;
+ return 1;
+}
+#endif
+
+#if DOXYGEN || FAT_WRITE_SUPPORT
+/**
+ * \ingroup fat_dir
+ * Creates a directory.
+ *
+ * Creates a directory and obtains its directory entry.
+ * If the directory to create already exists, its
+ * directory entry will be returned within the dir_entry
+ * parameter.
+ *
+ * \note The notes which apply to fat_create_file() also
+ * apply to this function.
+ *
+ * \param[in] parent The handle of the parent directory of the new directory.
+ * \param[in] dir The name of the directory to create.
+ * \param[out] dir_entry The directory entry to fill for the new directory.
+ * \returns 0 on failure, 1 on success.
+ * \see fat_delete_dir
+ */
+uint8_t fat_create_dir(struct fat_dir_struct* parent, const char* dir, struct fat_dir_entry_struct* dir_entry)
+{
+ if(!parent || !dir || !dir[0] || !dir_entry)
+ return 0;
+
+ /* check if the file or directory already exists */
+ while(fat_read_dir(parent, dir_entry))
+ {
+ if(strcmp(dir, dir_entry->long_name) == 0)
+ {
+ fat_reset_dir(parent);
+ return 0;
+ }
+ }
+
+ struct fat_fs_struct* fs = parent->fs;
+
+ /* allocate cluster which will hold directory entries */
+ cluster_t dir_cluster = fat_append_clusters(fs, 0, 1);
+ if(!dir_cluster)
+ return 0;
+
+ /* clear cluster to prevent bogus directory entries */
+ fat_clear_cluster(fs, dir_cluster);
+
+ memset(dir_entry, 0, sizeof(*dir_entry));
+ dir_entry->attributes = FAT_ATTRIB_DIR;
+
+ /* create "." directory self reference */
+ dir_entry->entry_offset = fs->header.cluster_zero_offset +
+ (offset_t) (dir_cluster - 2) * fs->header.cluster_size;
+ dir_entry->long_name[0] = '.';
+ dir_entry->cluster = dir_cluster;
+ if(!fat_write_dir_entry(fs, dir_entry))
+ {
+ fat_free_clusters(fs, dir_cluster);
+ return 0;
+ }
+
+ /* create ".." parent directory reference */
+ dir_entry->entry_offset += 32;
+ dir_entry->long_name[1] = '.';
+ dir_entry->cluster = parent->dir_entry.cluster;
+ if(!fat_write_dir_entry(fs, dir_entry))
+ {
+ fat_free_clusters(fs, dir_cluster);
+ return 0;
+ }
+
+ /* fill directory entry */
+ strncpy(dir_entry->long_name, dir, sizeof(dir_entry->long_name) - 1);
+ dir_entry->cluster = dir_cluster;
+
+ /* find place where to store directory entry */
+ if(!(dir_entry->entry_offset = fat_find_offset_for_dir_entry(fs, parent, dir_entry)))
+ {
+ fat_free_clusters(fs, dir_cluster);
+ return 0;
+ }
+
+ /* write directory to disk */
+ if(!fat_write_dir_entry(fs, dir_entry))
+ {
+ fat_free_clusters(fs, dir_cluster);
+ return 0;
+ }
+
+ return 1;
+}
+#endif
+
+/**
+ * \ingroup fat_dir
+ * Deletes a directory.
+ *
+ * This is just a synonym for fat_delete_file().
+ * If a directory is deleted without first deleting its
+ * subdirectories and files, disk space occupied by these
+ * files will get wasted as there is no chance to release
+ * it and mark it as free.
+ *
+ * \param[in] fs The filesystem on which to operate.
+ * \param[in] dir_entry The directory entry of the directory to delete.
+ * \returns 0 on failure, 1 on success.
+ * \see fat_create_dir
+ */
+#ifdef DOXYGEN
+uint8_t fat_delete_dir(struct fat_fs_struct* fs, struct fat_dir_entry_struct* dir_entry);
+#endif
+
+/**
+ * \ingroup fat_dir
+ * Moves or renames a directory.
+ *
+ * This is just a synonym for fat_move_file().
+ *
+ * \param[in] fs The filesystem on which to operate.
+ * \param[in,out] dir_entry The directory entry of the directory to move.
+ * \param[in] parent_new The handle of the new parent directory.
+ * \param[in] dir_new The directory's new name.
+ * \returns 0 on failure, 1 on success.
+ * \see fat_create_dir, fat_delete_dir, fat_move_file
+ */
+#ifdef DOXYGEN
+uint8_t fat_move_dir(struct fat_fs_struct* fs, struct fat_dir_entry_struct* dir_entry, struct fat_dir_struct* parent_new, const char* dir_new);
+#endif
+
+#if DOXYGEN || FAT_DATETIME_SUPPORT
+/**
+ * \ingroup fat_file
+ * Returns the modification date of a file.
+ *
+ * \param[in] dir_entry The directory entry of which to return the modification date.
+ * \param[out] year The year the file was last modified.
+ * \param[out] month The month the file was last modified.
+ * \param[out] day The day the file was last modified.
+ */
+void fat_get_file_modification_date(const struct fat_dir_entry_struct* dir_entry, uint16_t* year, uint8_t* month, uint8_t* day)
+{
+ if(!dir_entry)
+ return;
+
+ *year = 1980 + ((dir_entry->modification_date >> 9) & 0x7f);
+ *month = (dir_entry->modification_date >> 5) & 0x0f;
+ *day = (dir_entry->modification_date >> 0) & 0x1f;
+}
+#endif
+
+#if DOXYGEN || FAT_DATETIME_SUPPORT
+/**
+ * \ingroup fat_file
+ * Returns the modification time of a file.
+ *
+ * \param[in] dir_entry The directory entry of which to return the modification time.
+ * \param[out] hour The hour the file was last modified.
+ * \param[out] min The min the file was last modified.
+ * \param[out] sec The sec the file was last modified.
+ */
+void fat_get_file_modification_time(const struct fat_dir_entry_struct* dir_entry, uint8_t* hour, uint8_t* min, uint8_t* sec)
+{
+ if(!dir_entry)
+ return;
+
+ *hour = (dir_entry->modification_time >> 11) & 0x1f;
+ *min = (dir_entry->modification_time >> 5) & 0x3f;
+ *sec = ((dir_entry->modification_time >> 0) & 0x1f) * 2;
+}
+#endif
+
+#if DOXYGEN || (FAT_WRITE_SUPPORT && FAT_DATETIME_SUPPORT)
+/**
+ * \ingroup fat_file
+ * Sets the modification time of a date.
+ *
+ * \param[in] dir_entry The directory entry for which to set the modification date.
+ * \param[in] year The year the file was last modified.
+ * \param[in] month The month the file was last modified.
+ * \param[in] day The day the file was last modified.
+ */
+void fat_set_file_modification_date(struct fat_dir_entry_struct* dir_entry, uint16_t year, uint8_t month, uint8_t day)
+{
+ if(!dir_entry)
+ return;
+
+ dir_entry->modification_date =
+ ((year - 1980) << 9) |
+ ((uint16_t) month << 5) |
+ ((uint16_t) day << 0);
+}
+#endif
+
+#if DOXYGEN || (FAT_WRITE_SUPPORT && FAT_DATETIME_SUPPORT)
+/**
+ * \ingroup fat_file
+ * Sets the modification time of a file.
+ *
+ * \param[in] dir_entry The directory entry for which to set the modification time.
+ * \param[in] hour The year the file was last modified.
+ * \param[in] min The month the file was last modified.
+ * \param[in] sec The day the file was last modified.
+ */
+void fat_set_file_modification_time(struct fat_dir_entry_struct* dir_entry, uint8_t hour, uint8_t min, uint8_t sec)
+{
+ if(!dir_entry)
+ return;
+
+ dir_entry->modification_time =
+ ((uint16_t) hour << 11) |
+ ((uint16_t) min << 5) |
+ ((uint16_t) sec >> 1) ;
+}
+#endif
+
+/**
+ * \ingroup fat_fs
+ * Returns the amount of total storage capacity of the filesystem in bytes.
+ *
+ * \param[in] fs The filesystem on which to operate.
+ * \returns 0 on failure, the filesystem size in bytes otherwise.
+ */
+offset_t fat_get_fs_size(const struct fat_fs_struct* fs)
+{
+ if(!fs)
+ return 0;
+
+#if FAT_FAT32_SUPPORT
+ if(fs->partition->type == PARTITION_TYPE_FAT32)
+ return (offset_t) (fs->header.fat_size / 4 - 2) * fs->header.cluster_size;
+ else
+#endif
+ return (offset_t) (fs->header.fat_size / 2 - 2) * fs->header.cluster_size;
+}
+
+/**
+ * \ingroup fat_fs
+ * Returns the amount of free storage capacity on the filesystem in bytes.
+ *
+ * \note As the FAT filesystem is cluster based, this function does not
+ * return continuous values but multiples of the cluster size.
+ *
+ * \param[in] fs The filesystem on which to operate.
+ * \returns 0 on failure, the free filesystem space in bytes otherwise.
+ */
+offset_t fat_get_fs_free(const struct fat_fs_struct* fs)
+{
+ if(!fs)
+ return 0;
+
+ uint8_t fat[32];
+ struct fat_usage_count_callback_arg count_arg;
+ count_arg.cluster_count = 0;
+ count_arg.buffer_size = sizeof(fat);
+
+ offset_t fat_offset = fs->header.fat_offset;
+ uint32_t fat_size = fs->header.fat_size;
+ while(fat_size > 0)
+ {
+ uintptr_t length = UINTPTR_MAX - 1;
+ if(fat_size < length)
+ length = fat_size;
+
+ if(!fs->partition->device_read_interval(fat_offset,
+ fat,
+ sizeof(fat),
+ length,
+#if FAT_FAT32_SUPPORT
+ (fs->partition->type == PARTITION_TYPE_FAT16) ?
+ fat_get_fs_free_16_callback :
+ fat_get_fs_free_32_callback,
+#else
+ fat_get_fs_free_16_callback,
+#endif
+ &count_arg
+ )
+ )
+ return 0;
+
+ fat_offset += length;
+ fat_size -= length;
+ }
+
+ return (offset_t) count_arg.cluster_count * fs->header.cluster_size;
+}
+
+/**
+ * \ingroup fat_fs
+ * Callback function used for counting free clusters in a FAT.
+ */
+uint8_t fat_get_fs_free_16_callback(uint8_t* buffer, offset_t offset, void* p)
+{
+ struct fat_usage_count_callback_arg* count_arg = (struct fat_usage_count_callback_arg*) p;
+ uintptr_t buffer_size = count_arg->buffer_size;
+
+ for(uintptr_t i = 0; i < buffer_size; i += 2, buffer += 2)
+ {
+ uint16_t cluster = read16(buffer);
+ if(cluster == HTOL16(FAT16_CLUSTER_FREE))
+ ++(count_arg->cluster_count);
+ }
+
+ return 1;
+}
+
+#if DOXYGEN || FAT_FAT32_SUPPORT
+/**
+ * \ingroup fat_fs
+ * Callback function used for counting free clusters in a FAT32.
+ */
+uint8_t fat_get_fs_free_32_callback(uint8_t* buffer, offset_t offset, void* p)
+{
+ struct fat_usage_count_callback_arg* count_arg = (struct fat_usage_count_callback_arg*) p;
+ uintptr_t buffer_size = count_arg->buffer_size;
+
+ for(uintptr_t i = 0; i < buffer_size; i += 4, buffer += 4)
+ {
+ uint32_t cluster = read32(buffer);
+ if(cluster == HTOL32(FAT32_CLUSTER_FREE))
+ ++(count_arg->cluster_count);
+ }
+
+ return 1;
+}
+#endif
+
git.droids-corp.org / protos / imu.git / commitdiff