nrf: Add micro:bit filesystem.

* ports/nrf: Add micro:bit filesystem.

This filesystem has been copied from BBC micro:bit sources [1] and
modified to work with the nRF5x port.

[1]: https://github.com/bbcmicrobit/micropython/blob/master/source/microbit/filesystem.c

* ports/nrf/modules/uos: Make listdir() and ilistdir() consistent.

This removes the optional direcotry paramter from ilistdir(). This is
not consistent with VFS, but makes more sense when using only the
microbit filesystem.

Saves about 100 bytes.

* ports/nrf/modules/uos: Add code size comment.

ports/nrf/Makefile

@@ -161,6 +161,7 @@ DRIVERS_SRC_C += $(addprefix modules/,\
 	machine/led.c \
 	machine/temp.c \
 	uos/moduos.c \
+	uos/microbitfs.c \
 	utime/modutime.c \
 	pyb/modpyb.c \
 	ubluepy/modubluepy.c \

ports/nrf/main.c

@@ -43,6 +43,7 @@
 #include "gccollect.h"
 #include "modmachine.h"
 #include "modmusic.h"
+#include "modules/uos/microbitfs.h"
 #include "led.h"
 #include "uart.h"
 #include "nrf.h"
@@ -138,6 +139,10 @@ soft_reset:
 
 pin_init0();
 
+#if MICROPY_HW_HAS_BUILTIN_FLASH
+    microbit_filesystem_init();
+#endif
+
 #if MICROPY_HW_HAS_SDCARD
     // if an SD card is present then mount it on /sd/
     if (sdcard_is_present()) {
@@ -215,6 +220,23 @@ pin_init0();
 }
 
 #if !MICROPY_VFS
+#if MICROPY_HW_HAS_BUILTIN_FLASH
+// Use micro:bit filesystem
+mp_lexer_t *mp_lexer_new_from_file(const char *filename) {
+    return uos_mbfs_new_reader(filename);
+}
+
+mp_import_stat_t mp_import_stat(const char *path) {
+    return uos_mbfs_import_stat(path);
+}
+
+STATIC mp_obj_t mp_builtin_open(size_t n_args, const mp_obj_t *args) {
+    return uos_mbfs_open(n_args, args);
+}
+MP_DEFINE_CONST_FUN_OBJ_VAR_BETWEEN(mp_builtin_open_obj, 1, 2, mp_builtin_open);
+
+#else
+// use dummy functions - no filesystem available
 mp_lexer_t *mp_lexer_new_from_file(const char *filename) {
     mp_raise_OSError(MP_ENOENT);
 }
@@ -228,6 +250,7 @@ STATIC mp_obj_t mp_builtin_open(size_t n_args, const mp_obj_t *args, mp_map_t *k
 }
 MP_DEFINE_CONST_FUN_OBJ_KW(mp_builtin_open_obj, 1, mp_builtin_open);
 #endif
+#endif
 
 void HardFault_Handler(void)
 {

ports/nrf/modules/uos/microbitfs.c

@@ -0,0 +1,714 @@
+/*
+ * This file is part of the Micro Python project, http://micropython.org/
+ *
+ * The MIT License (MIT)
+ *
+ * Copyright (c) 2016 Mark Shannon
+ * Copyright (c) 2017 Ayke van Laethem
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+
+#include <string.h>
+#include <stdio.h>
+#include <errno.h>
+#include <sys/stat.h>
+
+#include "microbitfs.h"
+#include "hal/hal_nvmc.h"
+#include "hal/hal_rng.h"
+#include "py/nlr.h"
+#include "py/obj.h"
+#include "py/stream.h"
+#include "py/runtime.h"
+#include "extmod/vfs.h"
+#include "mpconfigport.h"
+
+#if MICROPY_HW_HAS_BUILTIN_FLASH
+
+#define DEBUG_FILE 0
+#if DEBUG_FILE
+#define DEBUG(s) printf s
+#else
+#define DEBUG(s) (void)0
+#endif
+
+/**  How it works:
+ * The File System consists of up to MAX_CHUNKS_IN_FILE_SYSTEM chunks of CHUNK_SIZE each,
+ * plus one spare page which holds persistent configuration data and is used. for bulk erasing.
+ * The spare page is either the first or the last page and will be switched by a bulk erase.
+ * The exact number of chunks will depend on the amount of flash available.
+ *
+ * Each chunk consists of a one byte marker and a one byte tail
+ * The marker shows whether this chunk is the start of a file, the midst of a file
+ * (in which case it refers to the previous chunk in the file) or whether it is UNUSED
+ * (and erased) or FREED (which means it is unused, but not erased).
+ * Chunks are selected in a randomised round-robin fashion to even out wear on the flash
+ * memory as much as possible.
+ * A file consists of a linked list of chunks. The first chunk in a file contains its name
+ * as well as the end chunk and offset.
+ * Files are found by linear search of the chunks, this means that no meta-data needs to be stored
+ * outside of the file, which prevents wear hot-spots. Since there are fewer than 250 chunks,
+ * the search is fast enough.
+ *
+ * Chunks are numbered from 1 as we need to reserve 0 as the FREED marker.
+ *
+ * Writing to files relies on the persistent API which is high-level wrapper on top of the Nordic SDK.
+ */
+
+#define CHUNK_SIZE (1<<MBFS_LOG_CHUNK_SIZE)
+#define DATA_PER_CHUNK (CHUNK_SIZE-2)
+
+#define UNUSED_CHUNK 255
+#define FREED_CHUNK  0
+#define FILE_START 254
+#define PERSISTENT_DATA_MARKER 253
+
+/** Must be such that sizeof(file_header) < DATA_PER_CHUNK */
+#define MAX_FILENAME_LENGTH 120
+
+//Minimum number of free chunks to justify sweeping.
+//If this is too low it may cause excessive wear
+#define MIN_CHUNKS_FOR_SWEEP (HAL_NVMC_PAGESIZE / CHUNK_SIZE)
+
+#define FILE_NOT_FOUND ((uint8_t)-1)
+
+/** Maximum number of chunks allowed in filesystem. 252 chunks is 31.5kB */
+#define MAX_CHUNKS_IN_FILE_SYSTEM 252
+
+#define STATIC_ASSERT(e) extern char static_assert_failed[(e) ? 1 : -1]
+
+typedef struct _file_descriptor_obj {
+    mp_obj_base_t base;
+    uint8_t start_chunk;
+    uint8_t seek_chunk;
+    uint8_t seek_offset;
+    bool writable;
+    bool open;
+    bool binary;
+} file_descriptor_obj;
+
+typedef struct _file_header {
+    uint8_t end_offset;
+    uint8_t name_len;
+    char filename[MAX_FILENAME_LENGTH];
+} file_header;
+
+typedef struct _file_chunk {
+    uint8_t marker;
+    union {
+        char data[DATA_PER_CHUNK];
+        file_header header;
+    };
+    uint8_t next_chunk;
+} file_chunk;
+
+typedef struct _persistent_config_t {
+    // Must start with a marker, so that we can identify it.
+    uint8_t marker; // Should always be PERSISTENT_DATA_MARKER
+} persistent_config_t;
+
+extern const mp_obj_type_t uos_mbfs_fileio_type;
+extern const mp_obj_type_t uos_mbfs_textio_type;
+
+// Page indexes count down from the end of ROM.
+STATIC uint8_t first_page_index;
+STATIC uint8_t last_page_index;
+// The number of useable chunks in the file system.
+STATIC uint8_t chunks_in_file_system;
+// Index of chunk to start searches. This is randomised to even out wear.
+STATIC uint8_t start_index;
+STATIC file_chunk *file_system_chunks;
+
+// Defined by the linker
+extern byte _flash_user_start[];
+extern byte _flash_user_end[];
+
+STATIC_ASSERT((sizeof(file_chunk) == CHUNK_SIZE));
+
+// From micro:bit memory.h
+STATIC inline byte *rounddown(byte *addr, uint32_t align) {
+    return (byte*)(((uint32_t)addr)&(-align));
+}
+
+// From micro:bit memory.h
+STATIC inline byte *roundup(byte *addr, uint32_t align) {
+    return (byte*)((((uint32_t)addr)+align-1)&(-align));
+}
+
+
+STATIC inline void *first_page(void) {
+    return _flash_user_end - HAL_NVMC_PAGESIZE * first_page_index;
+}
+
+STATIC inline void *last_page(void) {
+    return _flash_user_end - HAL_NVMC_PAGESIZE * last_page_index;
+}
+
+STATIC void init_limits(void) {
+    // First determine where to end
+    byte *end = _flash_user_end;
+    end = rounddown(end, HAL_NVMC_PAGESIZE)-HAL_NVMC_PAGESIZE;
+    last_page_index = (_flash_user_end - end)/HAL_NVMC_PAGESIZE;
+
+    // Now find the start
+    byte *start = roundup(end - CHUNK_SIZE*MAX_CHUNKS_IN_FILE_SYSTEM, HAL_NVMC_PAGESIZE);
+    while (start < _flash_user_start) {
+        start += HAL_NVMC_PAGESIZE;
+    }
+    first_page_index = (_flash_user_end - start)/HAL_NVMC_PAGESIZE;
+    chunks_in_file_system = (end-start)>>MBFS_LOG_CHUNK_SIZE;
+}
+
+STATIC void randomise_start_index(void) {
+    start_index = hal_rng_generate() / (chunks_in_file_system-1) + 1;
+}
+
+void microbit_filesystem_init(void) {
+    init_limits();
+    randomise_start_index();
+    file_chunk *base = first_page();
+    if (base->marker == PERSISTENT_DATA_MARKER) {
+        file_system_chunks = &base[(HAL_NVMC_PAGESIZE>>MBFS_LOG_CHUNK_SIZE)-1];
+    } else if (((file_chunk *)last_page())->marker == PERSISTENT_DATA_MARKER) {
+        file_system_chunks = &base[-1];
+    } else {
+        hal_nvmc_write_byte(&((file_chunk *)last_page())->marker, PERSISTENT_DATA_MARKER);
+        file_system_chunks = &base[-1];
+    }
+}
+
+STATIC void copy_page(void *dest, void *src) {
+    DEBUG(("FILE DEBUG: Copying page from %lx to %lx.\r\n", (uint32_t)src, (uint32_t)dest));
+    hal_nvmc_erase_page((uint32_t)dest);
+    file_chunk *src_chunk = src;
+    file_chunk *dest_chunk = dest;
+    uint32_t chunks = HAL_NVMC_PAGESIZE>>MBFS_LOG_CHUNK_SIZE;
+    for (uint32_t i = 0; i < chunks; i++) {
+        if (src_chunk[i].marker != FREED_CHUNK) {
+            hal_nvmc_write_buffer(&dest_chunk[i], &src_chunk[i], CHUNK_SIZE);
+        }
+    }
+}
+
+// Move entire file system up or down one page, copying all used chunks
+// Freed chunks are not copied, so become erased.
+// There should be no erased chunks before the sweep (or it would be unnecessary)
+// but if there are this should work correctly.
+//
+// The direction of the sweep depends on whether the persistent data is in the first or last page
+// The persistent data is copied to RAM, leaving its page unused.
+// Then all the pages are copied, one by one, into the adjacent newly unused page.
+// Finally, the persistent data is saved back to the opposite end of the filesystem from whence it came.
+//
+STATIC void filesystem_sweep(void) {
+    persistent_config_t config;
+    uint8_t *page;
+    uint8_t *end_page;
+    int step;
+    uint32_t page_size = HAL_NVMC_PAGESIZE;
+    DEBUG(("FILE DEBUG: Sweeping file system\r\n"));
+    if (((file_chunk *)first_page())->marker == PERSISTENT_DATA_MARKER) {
+        config = *(persistent_config_t *)first_page();
+        page = first_page();
+        end_page = last_page();
+        step = page_size;
+    } else {
+        config = *(persistent_config_t *)last_page();
+        page = last_page();
+        end_page = first_page();
+        step = -page_size;
+    }
+    while (page != end_page) {
+        uint8_t *next_page = page+step;
+        hal_nvmc_erase_page((uint32_t)page);
+        copy_page(page, next_page);
+        page = next_page;
+    }
+    hal_nvmc_erase_page((uint32_t)end_page);
+    hal_nvmc_write_buffer(end_page, &config, sizeof(config));
+    microbit_filesystem_init();
+}
+
+
+STATIC inline byte *seek_address(file_descriptor_obj *self) {
+    return (byte*)&(file_system_chunks[self->seek_chunk].data[self->seek_offset]);
+}
+
+STATIC uint8_t microbit_find_file(const char *name, int name_len) {
+    for (uint8_t index = 1; index <= chunks_in_file_system; index++) {
+        const file_chunk *p = &file_system_chunks[index];
+        if (p->marker != FILE_START)
+            continue;
+        if (p->header.name_len != name_len)
+            continue;
+        if (memcmp(name, &p->header.filename[0], name_len) == 0) {
+            DEBUG(("FILE DEBUG: File found. index %d\r\n", index));
+            return index;
+        }
+    }
+    DEBUG(("FILE DEBUG: File not found.\r\n"));
+    return FILE_NOT_FOUND;
+}
+
+// Return a free, erased chunk.
+// Search the chunks:
+// 1  If an UNUSED chunk is found, then return that.
+// 2. If an entire page of FREED chunks is found, then erase the page and return the first chunk
+// 3. If the number of FREED chunks is >= MIN_CHUNKS_FOR_SWEEP, then
+// 3a. Sweep the filesystem and restart.
+// 3b. Fail and return FILE_NOT_FOUND
+//
+STATIC uint8_t find_chunk_and_erase(void) {
+    // Start search at a random chunk to spread the wear more evenly.
+    // Search for unused chunk
+    uint8_t index = start_index;
+    do {
+        const file_chunk *p = &file_system_chunks[index];
+        if (p->marker == UNUSED_CHUNK) {
+            DEBUG(("FILE DEBUG: Unused chunk found: %d\r\n", index));
+            return index;
+        }
+        index++;
+        if (index == chunks_in_file_system+1) index = 1;
+    } while (index != start_index);
+
+    // Search for FREED page, and total up FREED chunks
+    uint32_t freed_chunks = 0;
+    index = start_index;
+    uint32_t chunks_per_page = HAL_NVMC_PAGESIZE>>MBFS_LOG_CHUNK_SIZE;
+    do {
+        const file_chunk *p = &file_system_chunks[index];
+        if (p->marker == FREED_CHUNK) {
+            freed_chunks++;
+        }
+        if (HAL_NVMC_IS_PAGE_ALIGNED(p)) {
+            uint32_t i;
+            for (i = 0; i < chunks_per_page; i++) {
+                if (p[i].marker != FREED_CHUNK)
+                    break;
+            }
+            if (i == chunks_per_page) {
+                DEBUG(("FILE DEBUG: Found freed page of chunks: %d\r\n", index));
+                hal_nvmc_erase_page((uint32_t)&file_system_chunks[index]);
+                return index;
+            }
+        }
+        index++;
+        if (index == chunks_in_file_system+1) index = 1;
+    } while (index != start_index);
+    DEBUG(("FILE DEBUG: %lu free chunks\r\n", freed_chunks));
+    if (freed_chunks < MIN_CHUNKS_FOR_SWEEP) {
+        return FILE_NOT_FOUND;
+    }
+    // No freed pages, so sweep file system.
+    filesystem_sweep();
+    // This is guaranteed to succeed.
+    return find_chunk_and_erase();
+}
+
+STATIC mp_obj_t microbit_file_name(file_descriptor_obj *fd) {
+    return mp_obj_new_str(&(file_system_chunks[fd->start_chunk].header.filename[0]), file_system_chunks[fd->start_chunk].header.name_len, false);
+}
+
+STATIC file_descriptor_obj *microbit_file_descriptor_new(uint8_t start_chunk, bool write, bool binary);
+
+STATIC void clear_file(uint8_t chunk) {
+    do {
+        hal_nvmc_write_byte(&(file_system_chunks[chunk].marker), FREED_CHUNK);
+        DEBUG(("FILE DEBUG: Freeing chunk %d.\n", chunk));
+        chunk = file_system_chunks[chunk].next_chunk;
+    } while (chunk <= chunks_in_file_system);
+}
+
+STATIC file_descriptor_obj *microbit_file_open(const char *name, size_t name_len, bool write, bool binary) {
+    if (name_len > MAX_FILENAME_LENGTH) {
+        return NULL;
+    }
+    uint8_t index = microbit_find_file(name, name_len);
+    if (write) {
+        if (index != FILE_NOT_FOUND) {
+            // Free old file
+            clear_file(index);
+        }
+        index = find_chunk_and_erase();
+        if (index == FILE_NOT_FOUND) {
+            nlr_raise(mp_obj_new_exception_msg(&mp_type_OSError, "No more storage space"));
+        }
+        hal_nvmc_write_byte(&(file_system_chunks[index].marker), FILE_START);
+        hal_nvmc_write_byte(&(file_system_chunks[index].header.name_len), name_len);
+        hal_nvmc_write_buffer(&(file_system_chunks[index].header.filename[0]), name, name_len);
+    } else {
+        if (index == FILE_NOT_FOUND) {
+            return NULL;
+        }
+    }
+    return microbit_file_descriptor_new(index, write, binary);
+}
+
+STATIC file_descriptor_obj *microbit_file_descriptor_new(uint8_t start_chunk, bool write, bool binary) {
+    file_descriptor_obj *res = m_new_obj(file_descriptor_obj);
+    if (binary) {
+        res->base.type = &uos_mbfs_fileio_type;
+    } else {
+        res->base.type = &uos_mbfs_textio_type;
+    }
+    res->start_chunk = start_chunk;
+    res->seek_chunk = start_chunk;
+    res->seek_offset = file_system_chunks[start_chunk].header.name_len+2;
+    res->writable = write;
+    res->open = true;
+    res->binary = binary;
+    return res;
+}
+
+STATIC mp_obj_t microbit_remove(mp_obj_t filename) {
+    mp_uint_t name_len;
+    const char *name = mp_obj_str_get_data(filename, &name_len);
+    mp_uint_t index = microbit_find_file(name, name_len);
+    if (index == 255) {
+        nlr_raise(mp_obj_new_exception_msg(&mp_type_OSError, "file not found"));
+    }
+    clear_file(index);
+    return mp_const_none;
+}
+
+STATIC void check_file_open(file_descriptor_obj *self) {
+    if (!self->open) {
+        nlr_raise(mp_obj_new_exception_msg(&mp_type_ValueError, "I/O operation on closed file"));
+    }
+}
+
+STATIC int advance(file_descriptor_obj *self, uint32_t n, bool write) {
+    DEBUG(("FILE DEBUG: Advancing from chunk %d, offset %d.\r\n", self->seek_chunk, self->seek_offset));
+    self->seek_offset += n;
+    if (self->seek_offset == DATA_PER_CHUNK) {
+        self->seek_offset = 0;
+        if (write) {
+            uint8_t next_chunk = find_chunk_and_erase();
+            if (next_chunk == FILE_NOT_FOUND) {
+                clear_file(self->start_chunk);
+                self->open = false;
+                return ENOSPC;
+            }
+            // Link next chunk to this one
+            hal_nvmc_write_byte(&(file_system_chunks[self->seek_chunk].next_chunk), next_chunk);
+            hal_nvmc_write_byte(&(file_system_chunks[next_chunk].marker), self->seek_chunk);
+        }
+        self->seek_chunk = file_system_chunks[self->seek_chunk].next_chunk;
+    }
+    DEBUG(("FILE DEBUG: Advanced to chunk %d, offset %d.\r\n", self->seek_chunk, self->seek_offset));
+    return 0;
+}
+
+STATIC mp_uint_t microbit_file_read(mp_obj_t obj, void *buf, mp_uint_t size, int *errcode) {
+    file_descriptor_obj *self = (file_descriptor_obj *)obj;
+    check_file_open(self);
+    if (self->writable || file_system_chunks[self->start_chunk].marker == FREED_CHUNK) {
+        *errcode = EBADF;
+        return MP_STREAM_ERROR;
+    }
+    uint32_t bytes_read = 0;
+    uint8_t *data = buf;
+    while (1) {
+        mp_uint_t to_read = DATA_PER_CHUNK - self->seek_offset;
+        if (file_system_chunks[self->seek_chunk].next_chunk == UNUSED_CHUNK) {
+            uint8_t end_offset = file_system_chunks[self->start_chunk].header.end_offset;
+            if (end_offset == UNUSED_CHUNK) {
+                to_read = 0;
+            } else {
+                to_read = MIN(to_read, (mp_uint_t)end_offset-self->seek_offset);
+            }
+        }
+        to_read = MIN(to_read, size-bytes_read);
+        if (to_read == 0) {
+            break;
+        }
+        memcpy(data+bytes_read, seek_address(self), to_read);
+        advance(self, to_read, false);
+        bytes_read += to_read;
+    }
+    return bytes_read;
+}
+
+STATIC mp_uint_t microbit_file_write(mp_obj_t obj, const void *buf, mp_uint_t size, int *errcode) {
+    file_descriptor_obj *self = (file_descriptor_obj *)obj;
+    check_file_open(self);
+    if (!self->writable || file_system_chunks[self->start_chunk].marker == FREED_CHUNK) {
+        *errcode = EBADF;
+        return MP_STREAM_ERROR;
+    }
+    uint32_t len = size;
+    const uint8_t *data = buf;
+    while (len) {
+        uint32_t to_write = MIN(((uint32_t)(DATA_PER_CHUNK - self->seek_offset)), len);
+        hal_nvmc_write_buffer(seek_address(self), data, to_write);
+        int err = advance(self, to_write, true);
+        if (err) {
+            *errcode = err;
+            return MP_STREAM_ERROR;
+        }
+        data += to_write;
+        len -= to_write;
+    }
+    return size;
+}
+
+STATIC void microbit_file_close(file_descriptor_obj *fd) {
+    if (fd->writable) {
+        hal_nvmc_write_byte(&(file_system_chunks[fd->start_chunk].header.end_offset), fd->seek_offset);
+    }
+    fd->open = false;
+}
+
+STATIC mp_obj_t microbit_file_list(void) {
+    mp_obj_t res = mp_obj_new_list(0, NULL);
+    for (uint8_t index = 1; index <= chunks_in_file_system; index++) {
+        if (file_system_chunks[index].marker == FILE_START) {
+            mp_obj_t name = mp_obj_new_str(&file_system_chunks[index].header.filename[0], file_system_chunks[index].header.name_len, false);
+            mp_obj_list_append(res, name);
+        }
+    }
+    return res;
+}
+
+STATIC mp_obj_t microbit_file_size(mp_obj_t filename) {
+    mp_uint_t name_len;
+    const char *name = mp_obj_str_get_data(filename, &name_len);
+    uint8_t chunk = microbit_find_file(name, name_len);
+    if (chunk == 255) {
+        nlr_raise(mp_obj_new_exception_msg(&mp_type_OSError, "file not found"));
+    }
+    mp_uint_t len = 0;
+    uint8_t end_offset = file_system_chunks[chunk].header.end_offset;
+    uint8_t offset = file_system_chunks[chunk].header.name_len+2;
+    while (file_system_chunks[chunk].next_chunk != UNUSED_CHUNK) {
+        len += DATA_PER_CHUNK - offset;
+        chunk = file_system_chunks[chunk].next_chunk;
+        offset = 0;
+    }
+    len += end_offset - offset;
+    return mp_obj_new_int(len);
+}
+
+STATIC mp_uint_t file_read_byte(file_descriptor_obj *fd) {
+    if (file_system_chunks[fd->seek_chunk].next_chunk == UNUSED_CHUNK) {
+        uint8_t end_offset = file_system_chunks[fd->start_chunk].header.end_offset;
+        if (end_offset == UNUSED_CHUNK || fd->seek_offset == end_offset) {
+            return (mp_uint_t)-1;
+        }
+    }
+    mp_uint_t res = file_system_chunks[fd->seek_chunk].data[fd->seek_offset];
+    advance(fd, 1, false);
+    return res;
+}
+
+// Now follows the code to integrate this filesystem into the uos module.
+
+mp_lexer_t *uos_mbfs_new_reader(const char *filename) {
+    file_descriptor_obj *fd = microbit_file_open(filename, strlen(filename), false, false);
+    if (fd == NULL) {
+        mp_raise_OSError(MP_ENOENT);
+    }
+    mp_reader_t reader;
+    reader.data = fd;
+    reader.readbyte = (mp_uint_t(*)(void*))file_read_byte;
+    reader.close = (void(*)(void*))microbit_file_close; // no-op
+    return mp_lexer_new(qstr_from_str(filename), reader);
+}
+
+mp_import_stat_t uos_mbfs_import_stat(const char *path) {
+    uint8_t chunk = microbit_find_file(path, strlen(path));
+    if (chunk == FILE_NOT_FOUND) {
+        return MP_IMPORT_STAT_NO_EXIST;
+    } else {
+        return MP_IMPORT_STAT_FILE;
+    }
+}
+
+STATIC mp_obj_t uos_mbfs_file_name(mp_obj_t self) {
+    file_descriptor_obj *fd = (file_descriptor_obj*)self;
+    return microbit_file_name(fd);
+}
+STATIC MP_DEFINE_CONST_FUN_OBJ_1(uos_mbfs_file_name_obj, uos_mbfs_file_name);
+
+STATIC mp_obj_t uos_mbfs_file_close(mp_obj_t self) {
+    file_descriptor_obj *fd = (file_descriptor_obj*)self;
+    microbit_file_close(fd);
+    return mp_const_none;
+}
+STATIC MP_DEFINE_CONST_FUN_OBJ_1(uos_mbfs_file_close_obj, uos_mbfs_file_close);
+
+STATIC mp_obj_t uos_mbfs_mount(mp_obj_t self_in, mp_obj_t readonly, mp_obj_t mkfs) {
+    // This function is called only once (indirectly from main()) and is
+    // not exposed to Python code. So we can ignore the readonly flag and
+    // not care about mounting a second time.
+    microbit_filesystem_init();
+    return mp_const_none;
+}
+STATIC MP_DEFINE_CONST_FUN_OBJ_3(uos_mbfs_mount_obj, uos_mbfs_mount);
+
+STATIC mp_obj_t uos_mbfs_remove(mp_obj_t name) {
+    return microbit_remove(name);
+}
+MP_DEFINE_CONST_FUN_OBJ_1(uos_mbfs_remove_obj, uos_mbfs_remove);
+
+typedef struct {
+    mp_obj_base_t base;
+    mp_fun_1_t iternext;
+    uint8_t index;
+} uos_mbfs_ilistdir_it_t;
+
+STATIC mp_obj_t uos_mbfs_ilistdir_it_iternext(mp_obj_t self_in) {
+    uos_mbfs_ilistdir_it_t *self = MP_OBJ_TO_PTR(self_in);
+
+    // Read until the next FILE_START chunk.
+    for (; self->index <= chunks_in_file_system; self->index++) {
+        if (file_system_chunks[self->index].marker != FILE_START) {
+            continue;
+        }
+
+        // Get the file name as str object.
+        mp_obj_t name = mp_obj_new_str(&file_system_chunks[self->index].header.filename[0], file_system_chunks[self->index].header.name_len, false);
+
+        // make 3-tuple with info about this entry
+        mp_obj_tuple_t *t = MP_OBJ_TO_PTR(mp_obj_new_tuple(3, NULL));
+        t->items[0] = name;
+        t->items[1] = MP_OBJ_NEW_SMALL_INT(MP_S_IFREG); // all entries are files
+        t->items[2] = MP_OBJ_NEW_SMALL_INT(0); // no inode number
+
+        self->index++;
+        return MP_OBJ_FROM_PTR(t);
+    }
+
+    return MP_OBJ_STOP_ITERATION;
+}
+
+STATIC mp_obj_t uos_mbfs_ilistdir() {
+    uos_mbfs_ilistdir_it_t *iter = m_new_obj(uos_mbfs_ilistdir_it_t);
+    iter->base.type = &mp_type_polymorph_iter;
+    iter->iternext = uos_mbfs_ilistdir_it_iternext;
+    iter->index = 1;
+
+    return MP_OBJ_FROM_PTR(iter);
+}
+MP_DEFINE_CONST_FUN_OBJ_0(uos_mbfs_ilistdir_obj, uos_mbfs_ilistdir);
+
+MP_DEFINE_CONST_FUN_OBJ_0(uos_mbfs_listdir_obj, microbit_file_list);
+
+STATIC mp_obj_t microbit_file_writable(mp_obj_t self) {
+    return mp_obj_new_bool(((file_descriptor_obj *)self)->writable);
+}
+STATIC MP_DEFINE_CONST_FUN_OBJ_1(microbit_file_writable_obj, microbit_file_writable);
+
+STATIC const mp_map_elem_t uos_mbfs_file_locals_dict_table[] = {
+    { MP_OBJ_NEW_QSTR(MP_QSTR_close), (mp_obj_t)&uos_mbfs_file_close_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_name), (mp_obj_t)&uos_mbfs_file_name_obj },
+    //{ MP_ROM_QSTR(MP_QSTR___enter__), (mp_obj_t)&mp_identity_obj },
+    //{ MP_ROM_QSTR(MP_QSTR___exit__), (mp_obj_t)&file___exit___obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_writable), (mp_obj_t)&microbit_file_writable_obj },
+    /* Stream methods */
+    { MP_OBJ_NEW_QSTR(MP_QSTR_read), (mp_obj_t)&mp_stream_read_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_readinto), (mp_obj_t)&mp_stream_readinto_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_readline), (mp_obj_t)&mp_stream_unbuffered_readline_obj},
+    { MP_OBJ_NEW_QSTR(MP_QSTR_write), (mp_obj_t)&mp_stream_write_obj},
+};
+STATIC MP_DEFINE_CONST_DICT(uos_mbfs_file_locals_dict, uos_mbfs_file_locals_dict_table);
+
+
+STATIC const mp_stream_p_t textio_stream_p = {
+    .read = microbit_file_read,
+    .write = microbit_file_write,
+    .is_text = true,
+};
+
+const mp_obj_type_t uos_mbfs_textio_type = {
+    { &mp_type_type },
+    .name = MP_QSTR_TextIO,
+    .protocol = &textio_stream_p,
+    .locals_dict = (mp_obj_dict_t*)&uos_mbfs_file_locals_dict,
+};
+
+
+STATIC const mp_stream_p_t fileio_stream_p = {
+    .read = microbit_file_read,
+    .write = microbit_file_write,
+};
+
+const mp_obj_type_t uos_mbfs_fileio_type = {
+    { &mp_type_type },
+    .name = MP_QSTR_FileIO,
+    .protocol = &fileio_stream_p,
+    .locals_dict = (mp_obj_dict_t*)&uos_mbfs_file_locals_dict,
+};
+
+// From micro:bit fileobj.c
+mp_obj_t uos_mbfs_open(size_t n_args, const mp_obj_t *args) {
+    /// -1 means default; 0 explicitly false; 1 explicitly true.
+    int read = -1;
+    int text = -1;
+    if (n_args == 2) {
+        mp_uint_t len;
+        const char *mode = mp_obj_str_get_data(args[1], &len);
+        for (mp_uint_t i = 0; i < len; i++) {
+            if (mode[i] == 'r' || mode[i] == 'w') {
+                if (read >= 0) {
+                    goto mode_error;
+                }
+                read = (mode[i] == 'r');
+            } else if (mode[i] == 'b' || mode[i] == 't') {
+                if (text >= 0) {
+                    goto mode_error;
+                }
+                text = (mode[i] == 't');
+            } else {
+                goto mode_error;
+            }
+        }
+    }
+    mp_uint_t name_len;
+    const char *filename = mp_obj_str_get_data(args[0], &name_len);
+    file_descriptor_obj *res = microbit_file_open(filename, name_len, read == 0, text == 0);
+    if (res == NULL) {
+        nlr_raise(mp_obj_new_exception_msg(&mp_type_OSError, "file not found"));
+    }
+    return res;
+mode_error:
+    nlr_raise(mp_obj_new_exception_msg(&mp_type_ValueError, "illegal mode"));
+}
+
+STATIC mp_obj_t uos_mbfs_stat(mp_obj_t filename) {
+    mp_obj_t file_size = microbit_file_size(filename);
+
+    mp_obj_tuple_t *t = MP_OBJ_TO_PTR(mp_obj_new_tuple(10, NULL));
+    t->items[0] = MP_OBJ_NEW_SMALL_INT(MP_S_IFREG); // st_mode
+    t->items[1] = MP_OBJ_NEW_SMALL_INT(0); // st_ino
+    t->items[2] = MP_OBJ_NEW_SMALL_INT(0); // st_dev
+    t->items[3] = MP_OBJ_NEW_SMALL_INT(0); // st_nlink
+    t->items[4] = MP_OBJ_NEW_SMALL_INT(0); // st_uid
+    t->items[5] = MP_OBJ_NEW_SMALL_INT(0); // st_gid
+    t->items[6] = file_size;               // st_size
+    t->items[7] = MP_OBJ_NEW_SMALL_INT(0); // st_atime
+    t->items[8] = MP_OBJ_NEW_SMALL_INT(0); // st_mtime
+    t->items[9] = MP_OBJ_NEW_SMALL_INT(0); // st_ctime
+    return MP_OBJ_FROM_PTR(t);
+}
+MP_DEFINE_CONST_FUN_OBJ_1(uos_mbfs_stat_obj, uos_mbfs_stat);
+
+#endif // MICROPY_HW_HAS_BUILTIN_FLASH

ports/nrf/modules/uos/microbitfs.h

@@ -0,0 +1,52 @@
+/*
+ * This file is part of the Micro Python project, http://micropython.org/
+ *
+ * The MIT License (MIT)
+ *
+ * Copyright (c) 2016 Mark Shannon
+ * Copyright (c) 2017 Ayke van Laethem
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+#ifndef __MICROPY_INCLUDED_FILESYSTEM_H__
+#define __MICROPY_INCLUDED_FILESYSTEM_H__
+
+#include "py/obj.h"
+#include "py/lexer.h"
+
+#ifndef MBFS_LOG_CHUNK_SIZE
+// This property can be tuned to make the filesystem bigger (while keeping
+// the max number of blocks). Note that it cannot (currently) be increased
+// beyond 8 or uint8_t integers will overflow.
+// 2^7 == 128 bytes
+// (128-2) bytes/block * 252 blocks = 31752 usable bytes
+#define MBFS_LOG_CHUNK_SIZE 7
+#endif
+
+mp_obj_t uos_mbfs_open(size_t n_args, const mp_obj_t *args);
+void microbit_filesystem_init(void);
+mp_lexer_t *uos_mbfs_new_reader(const char *filename);
+mp_import_stat_t uos_mbfs_import_stat(const char *path);
+
+MP_DECLARE_CONST_FUN_OBJ_0(uos_mbfs_listdir_obj);
+MP_DECLARE_CONST_FUN_OBJ_0(uos_mbfs_ilistdir_obj);
+MP_DECLARE_CONST_FUN_OBJ_1(uos_mbfs_remove_obj);
+MP_DECLARE_CONST_FUN_OBJ_1(uos_mbfs_stat_obj);
+
+#endif // __MICROPY_INCLUDED_FILESYSTEM_H__

ports/nrf/modules/uos/moduos.c

@@ -33,6 +33,7 @@
 #include "py/objstr.h"
 #include "lib/oofatfs/ff.h"
 #include "lib/oofatfs/diskio.h"
+#include "modules/uos/microbitfs.h"
 #include "extmod/vfs.h"
 #include "extmod/vfs_fat.h"
 #include "genhdr/mpversion.h"
@@ -148,6 +149,12 @@ STATIC const mp_rom_map_elem_t os_module_globals_table[] = {
     { MP_ROM_QSTR(MP_QSTR_unlink), MP_ROM_PTR(&mp_vfs_remove_obj) }, // unlink aliases to remove
 
     { MP_ROM_QSTR(MP_QSTR_sync), MP_ROM_PTR(&mod_os_sync_obj) },
+
+#elif MICROPY_HW_HAS_BUILTIN_FLASH
+    { MP_ROM_QSTR(MP_QSTR_listdir), MP_ROM_PTR(&uos_mbfs_listdir_obj) },
+    { MP_ROM_QSTR(MP_QSTR_ilistdir), MP_ROM_PTR(&uos_mbfs_ilistdir_obj) }, // uses ~136 bytes
+    { MP_ROM_QSTR(MP_QSTR_stat), MP_ROM_PTR(&uos_mbfs_stat_obj) },         // uses ~228 bytes
+    { MP_ROM_QSTR(MP_QSTR_remove), MP_ROM_PTR(&uos_mbfs_remove_obj) },
 #endif
 
     /// \constant sep - separation character used in paths