/* * Copyright (c) 2008, Swedish Institute of Computer Science * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. Neither the name of the Institute nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE INSTITUTE AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE INSTITUTE OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * This file is part of the Contiki operating system. * */ /** * \file * Coffee: A flash file system for memory-contrained sensor systems. * \author * Nicolas Tsiftes */ #include "contiki-conf.h" #include "cfs/cfs.h" #include "cfs-coffee-arch.h" #include "cfs/cfs-coffee.h" #include "dev/watchdog.h" #include #include #define DEBUG 0 #if DEBUG #include #define PRINTF(...) printf(__VA_ARGS__) #else #define PRINTF(...) #endif #if COFFEE_PAGES_PER_SECTOR & (COFFEE_PAGES_PER_SECTOR - 1) #error COFFEE_PAGES_PER_SECTOR must be a power of two. #error Change COFFEE_PAGES_PER_SECTOR in cfs-coffee-arch.h. #endif #define kb * 1024UL #define Mb * (1024 kb) #define COFFEE_FD_FREE 0x0 #define COFFEE_FD_READ 0x1 #define COFFEE_FD_WRITE 0x2 #define COFFEE_FD_APPEND 0x4 #define COFFEE_FD_MODIFIED 0x8 #define FD_VALID(fd) \ ((fd) >= 0 && (fd) < COFFEE_FD_SET_SIZE && \ coffee_fd_set[(fd)].flags != COFFEE_FD_FREE) #define FD_READABLE(fd) (coffee_fd_set[(fd)].flags & CFS_READ) #define FD_WRITABLE(fd) (coffee_fd_set[(fd)].flags & CFS_WRITE) #define FD_APPENDABLE(fd) (coffee_fd_set[(fd)].flags & CFS_APPEND) #define FD_MODIFIED(fd) (coffee_fd_set[(fd)].flags & COFFEE_FD_MODIFIED) /* File header flags. */ #define COFFEE_FLAG_VALID 0x1 /* Completely written header. */ #define COFFEE_FLAG_ALLOCATED 0x2 /* Allocated file. */ #define COFFEE_FLAG_OBSOLETE 0x4 /* File marked for GC. */ #define COFFEE_FLAG_MODIFIED 0x8 /* Modified file, log exists. */ #define COFFEE_FLAG_LOG 0x10 /* Log file. */ #define COFFEE_FLAG_ISOLATED 0x20 /* Isolated page. */ #define COFFEE_FILE_ALLOCATED(hdr) ((hdr).flags & COFFEE_FLAG_ALLOCATED) #define COFFEE_FILE_FREE(hdr) !COFFEE_FILE_ALLOCATED(hdr) #define COFFEE_FILE_LOG(hdr) ((hdr).flags & COFFEE_FLAG_LOG) #define COFFEE_FILE_MODIFIED(hdr) ((hdr).flags & COFFEE_FLAG_MODIFIED) #define COFFEE_FILE_ISOLATED(hdr) ((hdr).flags & COFFEE_FLAG_ISOLATED) #define COFFEE_FILE_OBSOLETE(hdr) ((hdr).flags & COFFEE_FLAG_OBSOLETE) #define COFFEE_FILE_ACTIVE(hdr) (COFFEE_FILE_ALLOCATED(hdr) && \ !COFFEE_FILE_OBSOLETE(hdr) && \ !COFFEE_FILE_ISOLATED(hdr)) #define COFFEE_PAGE_COUNT (COFFEE_SIZE / COFFEE_PAGE_SIZE) #define COFFEE_PAGES_PER_SECTOR (COFFEE_SECTOR_SIZE / COFFEE_PAGE_SIZE) struct file_desc { coffee_offset_t offset; coffee_offset_t end; coffee_page_t file_page; coffee_page_t max_pages; int16_t next_log_record; uint8_t flags; }; struct dir_cache { coffee_offset_t page; char filename_start; }; struct file_header { coffee_page_t log_page; uint16_t log_records; uint16_t log_record_size; coffee_page_t max_pages; uint8_t eof_hint; uint8_t flags; char name[COFFEE_NAME_LENGTH]; } __attribute__((packed)); /* This is needed because of a buggy compiler. */ struct log_param { coffee_offset_t offset; const char *buf; uint16_t size; }; #define ABS_OFFSET(file_page, file_offset) \ ((file_page) * COFFEE_PAGE_SIZE + \ sizeof(struct file_header) + (file_offset)) static struct dir_cache dir_cache[COFFEE_DIR_CACHE_ENTRIES]; static struct file_desc coffee_fd_set[COFFEE_FD_SET_SIZE]; /*---------------------------------------------------------------------------*/ static void write_header(struct file_header *hdr, coffee_page_t page) { COFFEE_WRITE(hdr, sizeof(*hdr), page * COFFEE_PAGE_SIZE); } /*---------------------------------------------------------------------------*/ static void read_header(struct file_header *hdr, coffee_page_t page) { COFFEE_READ(hdr, sizeof(*hdr), page * COFFEE_PAGE_SIZE); } /*---------------------------------------------------------------------------*/ static void get_sector_status(uint16_t sector, coffee_page_t *active, coffee_page_t *free, coffee_page_t *obsolete) { static coffee_page_t skip_pages; static int last_pages_are_active; coffee_offset_t offset, sector_start; coffee_offset_t end; struct file_header hdr; coffee_page_t jump; int i; *active = *free = *obsolete = 0; if(sector == 0) { skip_pages = 0; last_pages_are_active = 0; } else if(skip_pages > COFFEE_PAGES_PER_SECTOR) { skip_pages -= COFFEE_PAGES_PER_SECTOR; if(last_pages_are_active) { *active = COFFEE_PAGES_PER_SECTOR; } else { *obsolete = COFFEE_PAGES_PER_SECTOR; } return; } sector_start = sector * COFFEE_SECTOR_SIZE; if(last_pages_are_active) { *active = skip_pages; } else { *obsolete = skip_pages; /* Split an obsolete file starting in the previous sector and mark the following pages as isolated. */ offset = sector_start; for(i = 0; i < skip_pages; i++) { COFFEE_READ(&hdr, sizeof(hdr), offset); hdr.flags |= COFFEE_FLAG_ISOLATED; COFFEE_WRITE(&hdr, sizeof(hdr), offset); offset += COFFEE_PAGE_SIZE; } PRINTF("Coffee: Isolated %u pages starting in sector %d\n", (unsigned)skip_pages, (int)sector); } offset = sector_start + skip_pages * COFFEE_PAGE_SIZE; end = (sector + 1) * COFFEE_SECTOR_SIZE; jump = 0; while(offset < end) { COFFEE_READ(&hdr, sizeof(hdr), offset); last_pages_are_active = 0; if(COFFEE_FILE_ACTIVE(hdr)) { last_pages_are_active = 1; jump = hdr.max_pages; *active += jump; } else if(COFFEE_FILE_ISOLATED(hdr)) { jump = 1; *obsolete++; } else if(COFFEE_FILE_OBSOLETE(hdr)) { jump = hdr.max_pages; *obsolete += jump; } else if(COFFEE_FILE_FREE(hdr)) { *free = (end - offset) / COFFEE_PAGE_SIZE; break; } offset += jump * COFFEE_PAGE_SIZE; } skip_pages = *active + *obsolete - COFFEE_PAGES_PER_SECTOR; if(skip_pages > 0) { if(last_pages_are_active) { *active = COFFEE_PAGES_PER_SECTOR - *obsolete; } else { *obsolete = COFFEE_PAGES_PER_SECTOR - *active; } } } /*---------------------------------------------------------------------------*/ static void dir_cache_add(char c, coffee_offset_t page) { int i; for(i = 0; i < COFFEE_DIR_CACHE_ENTRIES; i++) { if(dir_cache[i].filename_start == '\0') { dir_cache[i].filename_start = c; dir_cache[i].page = page; break; } } } /*---------------------------------------------------------------------------*/ static void dir_cache_del(coffee_offset_t page) { int i; for(i = 0; i < COFFEE_DIR_CACHE_ENTRIES; i++) { if(dir_cache[i].page == page) { dir_cache[i].filename_start = '\0'; } } } /*---------------------------------------------------------------------------*/ static coffee_offset_t dir_cache_find(const char *name) { int i; struct file_header hdr; for(i = 0; i < COFFEE_DIR_CACHE_ENTRIES; i++) { if(*name == dir_cache[i].filename_start) { read_header(&hdr, dir_cache[i].page); if(!COFFEE_FILE_ACTIVE(hdr)) { dir_cache[i].filename_start = '\0'; } else if(strcmp(hdr.name, name) == 0) { return dir_cache[i].page; } } } return -1; } /*---------------------------------------------------------------------------*/ static coffee_offset_t find_file(const char *name) { struct file_header hdr; coffee_page_t page; page = dir_cache_find(name); if(page >= 0) { return page; } page = 0; do { read_header(&hdr, page); if(COFFEE_FILE_ACTIVE(hdr)) { if(dir_cache_find(name) == -1) { dir_cache_add(name[0], page); } if(strcmp(name, hdr.name) == 0) { return page; } page += hdr.max_pages; } else if(COFFEE_FILE_ISOLATED(hdr)) { ++page; } else if(COFFEE_FILE_OBSOLETE(hdr)) { page += hdr.max_pages; } else { /* It follows from the properties of the page allocation algorithm that if a free page is encountered, then the rest of the sector is also free. */ page = (page + COFFEE_PAGES_PER_SECTOR) & ~(COFFEE_PAGES_PER_SECTOR - 1); } watchdog_periodic(); } while(page < COFFEE_PAGE_COUNT); return -1; } /*---------------------------------------------------------------------------*/ static void update_eof_hint(coffee_page_t file_page, coffee_offset_t end) { struct file_header hdr; coffee_page_t current_page; int part_size, i; uint8_t eof_hint; read_header(&hdr, file_page); current_page = (end + sizeof(hdr)) / COFFEE_PAGE_SIZE; part_size = hdr.max_pages / (sizeof(hdr.eof_hint) * CHAR_BIT); if(part_size == 0) { part_size = 1; } for(i = eof_hint = 0; i < sizeof(eof_hint) * CHAR_BIT; i++) { eof_hint |= (current_page >= (i + 1) * part_size) << i; } if(eof_hint > hdr.eof_hint) { hdr.eof_hint |= eof_hint; write_header(&hdr, file_page); } } /*---------------------------------------------------------------------------*/ static coffee_offset_t find_offset_in_file(int first_page) { struct file_header hdr; unsigned char buf[COFFEE_PAGE_SIZE]; int page; int i; int search_limit; coffee_offset_t range_start, range_end, part_size; read_header(&hdr, first_page); search_limit = 0; for(i = 0; i < sizeof(hdr.eof_hint) * CHAR_BIT; i++) { if(hdr.eof_hint >> i) { search_limit = i + 1; } } part_size = hdr.max_pages / sizeof(hdr.eof_hint) / CHAR_BIT; if(part_size == 0) { part_size = 1; } range_start = part_size * search_limit; range_end = range_start + part_size; if(range_end + 1 > hdr.max_pages) { range_end = hdr.max_pages - 1; } /* * Move from the end of the range towards the beginning and look for * a byte that has been modified. * * An important implication of this is that if the last written bytes * are zeroes, then these are skipped from the calculation. */ for(page = first_page + range_end; page >= first_page; page--) { watchdog_periodic(); COFFEE_READ(buf, sizeof(buf), page * COFFEE_PAGE_SIZE); for(i = COFFEE_PAGE_SIZE - 1; i >= 0; i--) { if(buf[i] != 0) { if(page == first_page) { return i < sizeof(hdr) ? 0 : 1 + i - sizeof(hdr); } else { return 1 + i + (page - first_page) * COFFEE_PAGE_SIZE - sizeof(hdr); } } } } /* All bytes are writable. */ return 0; } /*---------------------------------------------------------------------------*/ static int find_contiguous_pages(unsigned wanted) { int page, start; struct file_header hdr; start = -1; for(page = 0; page < COFFEE_PAGE_COUNT;) { read_header(&hdr, page); if(COFFEE_FILE_FREE(hdr)) { if(start == -1) { start = page; } else { if(start + wanted <= page) { return start; } } /* Jump to the next sector. */ page = (page + COFFEE_PAGES_PER_SECTOR) & ~(COFFEE_PAGES_PER_SECTOR - 1); } else if(COFFEE_FILE_ISOLATED(hdr)) { ++page; } else { start = -1; page += hdr.max_pages; } } return -1; } /*---------------------------------------------------------------------------*/ static int cfs_garbage_collect(void) { uint16_t sector; coffee_page_t active_pages, free_pages, obsolete_pages; int nerased; watchdog_stop(); PRINTF("Coffee: Running the file system garbage collector...\n"); /* * The garbage collector erases as many sectors as possible. A sector is * erasable if there are only free or obsolete pages in it. */ for(nerased = sector = 0; sector < COFFEE_SIZE / COFFEE_SECTOR_SIZE; sector++) { get_sector_status(sector, &active_pages, &free_pages, &obsolete_pages); PRINTF("Coffee: Sector %u has %u active, %u free, and %u obsolete pages.\n", sector, (unsigned)active_pages, (unsigned)free_pages, (unsigned)obsolete_pages); if(active_pages == 0 && obsolete_pages > 0) { COFFEE_ERASE(sector); nerased++; PRINTF("Coffee: Erased sector %d!\n", sector); } } watchdog_start(); return nerased; } /*---------------------------------------------------------------------------*/ static int remove_by_page(coffee_page_t page, int remove_log, int close_fds) { struct file_header hdr; int i; coffee_page_t log_page; if(page >= COFFEE_PAGE_COUNT) { return -1; } read_header(&hdr, page); if(!COFFEE_FILE_ACTIVE(hdr)) { return -1; } dir_cache_del(page); hdr.flags |= COFFEE_FLAG_OBSOLETE; write_header(&hdr, page); if(remove_log && COFFEE_FILE_MODIFIED(hdr)) { log_page = hdr.log_page; dir_cache_del(log_page); read_header(&hdr, log_page); hdr.flags |= COFFEE_FLAG_OBSOLETE; write_header(&hdr, log_page); } /* Close all file descriptors that reference the remove file. */ if(close_fds) { for(i = 0; i < COFFEE_FD_SET_SIZE; i++) { if(coffee_fd_set[i].file_page == page) { coffee_fd_set[i].flags = COFFEE_FD_FREE; } } } return 0; } /*---------------------------------------------------------------------------*/ static int reserve(const char *name, uint32_t size, int allow_duplicates) { struct file_header hdr; unsigned need_pages; int page; if(!allow_duplicates && find_file(name) >= 0) { return -1; } need_pages = (size + sizeof(hdr) + COFFEE_PAGE_SIZE - 1) / COFFEE_PAGE_SIZE; watchdog_stop(); page = find_contiguous_pages(need_pages); if(page < 0) { cfs_garbage_collect(); page = find_contiguous_pages(need_pages); if(page < 0) { watchdog_start(); return -1; } } memcpy(hdr.name, name, sizeof(hdr.name)); hdr.name[sizeof(hdr.name) - 1] = '\0'; hdr.max_pages = need_pages; hdr.flags = COFFEE_FLAG_ALLOCATED | COFFEE_FLAG_VALID; hdr.log_page = 0; hdr.eof_hint = 0; hdr.log_records = 0; hdr.log_record_size = 0; write_header(&hdr, page); PRINTF("Coffee: Reserved %u pages starting from %u for file %s\n", need_pages, page, name); dir_cache_add(name[0], page); watchdog_start(); return page; } /*---------------------------------------------------------------------------*/ static void adjust_log_config(struct file_header *hdr, uint16_t *log_record_size, uint16_t *log_records) { *log_record_size = hdr->log_record_size == 0 ? COFFEE_PAGE_SIZE : hdr->log_record_size; *log_records = hdr->log_records == 0 ? COFFEE_LOG_SIZE / *log_record_size : hdr->log_records; } /*---------------------------------------------------------------------------*/ static uint16_t modify_log_buffer(uint16_t log_record_size, coffee_offset_t *offset, uint16_t *size) { uint16_t region; region = *offset / log_record_size; *offset %= log_record_size; if(*size > log_record_size - *offset) { *size = log_record_size - *offset; } return region; } /*---------------------------------------------------------------------------*/ static int get_record_index(coffee_page_t log_page, uint16_t search_records, uint16_t region) { coffee_offset_t base; uint16_t processed; uint16_t batch_size; int16_t match_index, i; uint16_t record_count; base = ABS_OFFSET(log_page, sizeof(uint16_t) * search_records); record_count = search_records > COFFEE_LOG_TABLE_LIMIT ? COFFEE_LOG_TABLE_LIMIT : search_records; processed = 0; match_index = -1; { uint16_t indices[record_count]; while(processed < search_records && match_index < 0) { if(record_count + processed > search_records) { batch_size = search_records - processed; } else { batch_size = record_count; } base -= batch_size * sizeof(indices[0]); COFFEE_READ(&indices, sizeof(indices[0]) * batch_size, base); for(i = batch_size - 1; i >= 0; i--) { if(indices[i] - 1 == region) { match_index = search_records - processed - (batch_size - i); break; } } processed += batch_size; } } return match_index; } /*---------------------------------------------------------------------------*/ static int read_log_page(struct file_header *hdr, int16_t last_record, struct log_param *lp) { uint16_t region; int16_t match_index; uint16_t log_record_size; uint16_t log_records; coffee_offset_t base; uint16_t search_records; adjust_log_config(hdr, &log_record_size, &log_records); region = modify_log_buffer(log_record_size, &lp->offset, &lp->size); search_records = last_record < 0 ? log_records : last_record + 1; match_index = get_record_index(hdr->log_page, search_records, region); if(match_index < 0) { return -1; } base = hdr->log_page * COFFEE_PAGE_SIZE; base += sizeof(struct file_header) + log_records * sizeof(region); base += (coffee_offset_t)match_index * log_record_size; base += lp->offset; COFFEE_READ(lp->buf, lp->size, base); return lp->size; } /*---------------------------------------------------------------------------*/ static unsigned char * create_log_name(unsigned char *new, int max_size, unsigned char *old) { unsigned char suffix[] = ".log"; int len; len = strlen(old); if(len > max_size - sizeof(suffix)) { len = max_size - sizeof(suffix); } memcpy(new, old, len); memcpy(&new[len], suffix, sizeof(suffix)); return new; } /*---------------------------------------------------------------------------*/ static coffee_page_t create_log(coffee_page_t file_page, struct file_header *hdr) { coffee_page_t log_page; unsigned char log_name[sizeof(hdr->name)]; uint16_t log_record_size, log_records; coffee_offset_t size; int i; adjust_log_config(hdr, &log_record_size, &log_records); size = log_records * sizeof(uint16_t); /* Log index size. */ size += log_records * log_record_size; /* Log data size. */ log_page = reserve(create_log_name(log_name, sizeof(log_name), hdr->name), size, 0); if(log_page < 0) { return -1; } hdr->flags |= COFFEE_FLAG_MODIFIED; hdr->log_page = log_page; write_header(hdr, file_page); read_header(hdr, log_page); hdr->flags |= COFFEE_FLAG_LOG; write_header(hdr, log_page); for(i = 0; i < COFFEE_FD_SET_SIZE; i++) { if(coffee_fd_set[i].file_page == file_page && coffee_fd_set[i].flags != COFFEE_FD_FREE) { coffee_fd_set[i].flags |= COFFEE_FD_MODIFIED; } } return log_page; } /*---------------------------------------------------------------------------*/ static int merge_log(coffee_page_t file_page, int extend) { coffee_page_t log_page, new_file_page; struct file_header hdr, hdr2; int fd, n; coffee_offset_t offset; coffee_page_t max_pages; read_header(&hdr, file_page); log_page = hdr.log_page; fd = cfs_open(hdr.name, CFS_READ); if(fd < 0) { return -1; } /* * The reservation function adds extra space for the header, which has * already been calculated with in the previous reservation. */ max_pages = hdr.max_pages * (1 << extend); new_file_page = reserve(hdr.name, max_pages * COFFEE_PAGE_SIZE - sizeof(hdr), 1); if(new_file_page < 0) { cfs_close(fd); return -1; } offset = 0; do { char buf[hdr.log_record_size == 0 ? COFFEE_PAGE_SIZE : hdr.log_record_size]; watchdog_periodic(); n = cfs_read(fd, buf, sizeof(buf)); if(n < 0) { remove_by_page(new_file_page, 0, 0); cfs_close(fd); return -1; } else if(n > 0) { COFFEE_WRITE(buf, n, ABS_OFFSET(new_file_page, offset)); offset += n; } } while(n != 0); if(remove_by_page(file_page, 1, 0) < 0) { remove_by_page(new_file_page, 0, 0); cfs_close(fd); return -1; } /* Copy the log configuration and the EOF hint. */ read_header(&hdr2, new_file_page); hdr2.log_record_size = hdr.log_record_size; hdr2.log_records = hdr.log_records; hdr2.eof_hint = hdr.eof_hint; write_header(&hdr2, new_file_page); /* Point the file descriptors to the new file page. */ for(n = 0; n < COFFEE_FD_SET_SIZE; n++) { if(coffee_fd_set[n].file_page == file_page) { coffee_fd_set[n].file_page = new_file_page; coffee_fd_set[n].max_pages = max_pages; coffee_fd_set[n].flags &= ~COFFEE_FD_MODIFIED; coffee_fd_set[n].next_log_record = -1; } } cfs_close(fd); return 0; } /*---------------------------------------------------------------------------*/ static int find_next_record(struct file_desc *fdp, coffee_page_t log_page, int log_records) { int log_record, i, preferred_batch_size; preferred_batch_size = log_records > COFFEE_LOG_TABLE_LIMIT ? COFFEE_LOG_TABLE_LIMIT : log_records; if(fdp->next_log_record == 0) { /* The next log record is unknown. Search for it. */ uint16_t indices[preferred_batch_size]; uint16_t processed; uint16_t batch_size; log_record = log_records; for(processed = 0; processed < log_records;) { batch_size = log_records - processed >= preferred_batch_size ? preferred_batch_size : log_records - processed; COFFEE_READ(&indices, batch_size * sizeof(indices[0]), ABS_OFFSET(log_page, processed * sizeof(indices[0]))); for(i = 0; i < batch_size && indices[i] != 0; i++); log_record = i; if(log_record < batch_size) { log_record += processed; break; } processed += batch_size; } } else { log_record = fdp->next_log_record; } return log_record; } /*---------------------------------------------------------------------------*/ static int write_log_page(struct file_desc *fdp, struct log_param *lp) { struct file_header hdr; uint16_t region; coffee_page_t log_page; int16_t log_record; uint16_t log_record_size; uint16_t log_records; coffee_offset_t base; struct log_param lp_out; read_header(&hdr, fdp->file_page); adjust_log_config(&hdr, &log_record_size, &log_records); region = modify_log_buffer(log_record_size, &lp->offset, &lp->size); log_page = 0; if(COFFEE_FILE_MODIFIED(hdr)) { /* A log structure has already been created. */ log_page = hdr.log_page; log_record = find_next_record(fdp, log_page, log_records); if(log_record >= log_records) { /* The log is full; merge the log. */ PRINTF("Coffee: Merging the file %s with its log\n", hdr.name); update_eof_hint(fdp->file_page, fdp->end); return merge_log(fdp->file_page, 0); } } else { /* Create a log structure. */ log_page = create_log(fdp->file_page, &hdr); if(log_page < 0) { return -1; } PRINTF("Coffee: Created a log structure for file %s at page %u\n", hdr.name, (unsigned)log_page); hdr.log_page = log_page; log_record = 0; } { unsigned char copy_buf[log_record_size]; lp_out.offset = region * log_record_size; lp_out.buf = copy_buf; lp_out.size = log_record_size; if((lp->offset > 0 || lp->size != log_record_size) && read_log_page(&hdr, fdp->next_log_record - 1, &lp_out) < 0) { COFFEE_READ(copy_buf, sizeof(copy_buf), ABS_OFFSET(fdp->file_page, region * log_record_size)); } memcpy((char *) ©_buf + lp->offset, lp->buf, lp->size); base = (coffee_offset_t)log_page * COFFEE_PAGE_SIZE; base += sizeof(hdr); base += log_records * sizeof(region); base += (coffee_offset_t)log_record * log_record_size; COFFEE_WRITE(copy_buf, sizeof(copy_buf), base); ++region; COFFEE_WRITE(®ion, sizeof(region), ABS_OFFSET(log_page, log_record * sizeof(region))); fdp->next_log_record = log_record + 1; } return lp->size; } /*---------------------------------------------------------------------------*/ static int get_available_fd(void) { int i; for(i = 0; i < COFFEE_FD_SET_SIZE; i++) { if(coffee_fd_set[i].flags == COFFEE_FD_FREE) { return i; } } return -1; } /*---------------------------------------------------------------------------*/ int cfs_open(const char *name, int flags) { int fd; int page; struct file_header hdr; struct file_desc *fdp; fd = get_available_fd(); if(fd < 0) { PRINTF("Coffee: Failed to allocate a new file descriptor!\n"); return -1; } fdp = &coffee_fd_set[fd]; page = find_file(name); if(page < 0) { if((flags & (CFS_READ | CFS_WRITE)) == CFS_READ) { return -1; } if((page = reserve(name, COFFEE_DYN_SIZE, 1)) < 0) { return -1; } fdp->max_pages = (COFFEE_DYN_SIZE + sizeof(hdr) + COFFEE_PAGE_SIZE - 1) / COFFEE_PAGE_SIZE; } else { read_header(&hdr, page); if(COFFEE_FILE_MODIFIED(hdr)) { fdp->flags = COFFEE_FD_MODIFIED; } else fdp->max_pages = hdr.max_pages; } fdp->file_page = page; fdp->flags |= flags; fdp->end = find_offset_in_file(page); fdp->offset = flags & CFS_APPEND ? fdp->end : 0; fdp->next_log_record = 0; return fd; } /*---------------------------------------------------------------------------*/ void cfs_close(int fd) { if(FD_VALID(fd)) { update_eof_hint(coffee_fd_set[fd].file_page, coffee_fd_set[fd].end); coffee_fd_set[fd].flags = COFFEE_FD_FREE; } } /*---------------------------------------------------------------------------*/ unsigned cfs_seek(int fd, unsigned offset) { struct file_header hdr; if(!FD_VALID(fd)) { return -1; } read_header(&hdr, coffee_fd_set[fd].file_page); /* Check if the offset is within the file boundary. */ if(sizeof(hdr) + offset >= hdr.max_pages * COFFEE_PAGE_SIZE || sizeof(hdr) + offset < offset) { return -1; } if(coffee_fd_set[fd].end < offset) { coffee_fd_set[fd].end = offset; } return coffee_fd_set[fd].offset = offset; } /*---------------------------------------------------------------------------*/ int cfs_coffee_remove(const char *name) { int page; page = find_file(name); if(page < 0) { return -1; } return remove_by_page(page, 1, 1); } /*---------------------------------------------------------------------------*/ int cfs_read(int fd, void *buf, unsigned size) { struct file_header hdr; struct file_desc *fdp; unsigned remains, read_chunk; int r; coffee_offset_t base, offset; struct log_param lp; if(!(FD_VALID(fd) && FD_READABLE(fd))) { return -1; } fdp = &coffee_fd_set[fd]; if(fdp->offset + size > fdp->end) { size = fdp->end - fdp->offset; } read_chunk = COFFEE_PAGE_SIZE; if(FD_MODIFIED(fd)) { read_header(&hdr, fdp->file_page); if(hdr.log_record_size > 0) { read_chunk = hdr.log_record_size; } } remains = size; base = fdp->offset; offset = 0; /* * Fill the buffer by copying from the log in first hand, or the * ordinary file if the page has no log record. */ while(remains) { watchdog_periodic(); r = -1; if(FD_MODIFIED(fd)) { lp.offset = base + offset; lp.buf = (char *)buf + offset; lp.size = remains; r = read_log_page(&hdr, fdp->next_log_record - 1, &lp); } /* Read from the original file if we cannot find the data in the log. */ if(r < 0) { r = remains > read_chunk ? read_chunk : remains; COFFEE_READ((char *) buf + offset, r, ABS_OFFSET(fdp->file_page, base + offset)); } remains -= r; offset += r; } fdp->offset += offset; return offset; } /*---------------------------------------------------------------------------*/ int cfs_write(int fd, const void *buf, unsigned size) { struct file_desc *fdp; int i; struct log_param lp; coffee_offset_t remains; if(!(FD_VALID(fd) && FD_WRITABLE(fd))) { return -1; } fdp = &coffee_fd_set[fd]; /* Attempt to extend the file if we try to write past the end. */ while(size + fdp->offset + sizeof(struct file_header) > fdp->max_pages * COFFEE_PAGE_SIZE) { update_eof_hint(fdp->file_page, fdp->end); if(merge_log(fdp->file_page, 1) < 0) { return -1; } } if(fdp->offset < fdp->end) { remains = size; while(remains) { lp.offset = fdp->offset; lp.buf = (char *)buf + size - remains; lp.size = remains; i = write_log_page(fdp, &lp); if(i == 0) { /* The file was merged with the log. Try again. */ i = write_log_page(fdp, &lp); } if(i < 0) { return size - remains > 0 ? size - remains : -1; } remains -= i; fdp->offset += i; } } else { COFFEE_WRITE(buf, size, ABS_OFFSET(fdp->file_page, fdp->offset)); fdp->offset += size; } if(fdp->offset > fdp->end) { /* * The file has been extended and the file descriptors * referencing this file must be updated. */ for(i = 0; i < COFFEE_FD_SET_SIZE; i++) { if(coffee_fd_set[i].file_page == fdp->file_page) { coffee_fd_set[i].end = coffee_fd_set[i].offset; } } } return size; } /*---------------------------------------------------------------------------*/ int cfs_opendir(struct cfs_dir *dir, const char *name) { /* We have only a root directory. */ if(name[0] != '/' || name[1] != '\0') { return -1; } *(coffee_page_t *)dir->dummy_space = 0; return 0; } /*---------------------------------------------------------------------------*/ int cfs_readdir(struct cfs_dir *dir, struct cfs_dirent *record) { struct file_header hdr; coffee_page_t page; for(page = *(coffee_page_t *)dir->dummy_space; page < COFFEE_PAGE_COUNT;) { watchdog_periodic(); read_header(&hdr, page); if(COFFEE_FILE_FREE(hdr)) { page = (page + COFFEE_PAGES_PER_SECTOR) & ~(COFFEE_PAGES_PER_SECTOR - 1); } else if(COFFEE_FILE_ISOLATED(hdr)) { ++page; } else if(COFFEE_FILE_ACTIVE(hdr) && !COFFEE_FILE_LOG(hdr)) { memcpy(record->name, hdr.name, sizeof(record->name)); record->name[sizeof(record->name) - 1] = '\0'; record->size = find_offset_in_file(page); page += hdr.max_pages; *(coffee_page_t *)dir->dummy_space = page; return 0; } else { page += hdr.max_pages; } } return -1; } /*---------------------------------------------------------------------------*/ void cfs_closedir(struct cfs_dir *dir) { return; } /*---------------------------------------------------------------------------*/ int cfs_coffee_reserve(const char *name, uint32_t size) { return reserve(name, size, 0); } /*---------------------------------------------------------------------------*/ int cfs_coffee_configure_log(const char *file, unsigned log_size, unsigned log_record_size) { coffee_page_t page; struct file_header hdr; if(log_record_size == 0 || log_record_size > COFFEE_PAGE_SIZE || log_size < log_record_size) { return -1; } page = find_file(file); if(page < 0) { return -1; } read_header(&hdr, page); if(COFFEE_FILE_MODIFIED(hdr)) { /* Too late to customize the log. */ return -1; } hdr.log_records = log_size / log_record_size; hdr.log_record_size = log_record_size; write_header(&hdr, page); return 0; } /*---------------------------------------------------------------------------*/ int cfs_coffee_format(void) { int nsectors, i; nsectors = COFFEE_SIZE / COFFEE_SECTOR_SIZE; /* Reject format requests when the configuration is incorrect. */ if(nsectors < 1 || (COFFEE_START & (COFFEE_SECTOR_SIZE - 1))) { return -1; } PRINTF("Coffee: Formatting %d sectors", nsectors); watchdog_stop(); for(i = 0; i < nsectors; i++) { COFFEE_ERASE(i); PRINTF("."); } watchdog_start(); /* All file descriptors have become invalid. */ for(i = 0; i < COFFEE_FD_SET_SIZE; i++) { coffee_fd_set[i].flags = COFFEE_FD_FREE; } PRINTF("done!\n"); return 0; }