From 6015af28ee6d44d6c7b21f3844c90df9239f66f4 Mon Sep 17 00:00:00 2001 From: Patrick Delaunay Date: Mon, 14 Oct 2019 09:28:04 +0200 Subject: dfu: add backend for MTD device Add DFU backend for MTD device: allow to read and write on all MTD device (NAND, SPI-NOR, SPI-NAND,...) For example : > set dfu_alt_info "nand_raw raw 0x0 0x100000" > dfu 0 mtd nand0 This MTD backend provides the same level than dfu nand backend for NAND and dfu sf backend for SPI-NOR; So it can replace booth of them but it also add support of spi-nand. > set dfu_alt_info "nand_raw raw 0x0 0x100000" > dfu 0 mtd spi-nand0 The backend code is based on the "mtd" command introduced by commit 5db66b3aee6f ("cmd: mtd: add 'mtd' command") Signed-off-by: Patrick Delaunay --- drivers/dfu/Kconfig | 6 ++ drivers/dfu/Makefile | 1 + drivers/dfu/dfu.c | 5 +- drivers/dfu/dfu_mtd.c | 232 ++++++++++++++++++++++++++++++++++++++++++++++++++ 4 files changed, 243 insertions(+), 1 deletion(-) create mode 100644 drivers/dfu/dfu_mtd.c (limited to 'drivers') diff --git a/drivers/dfu/Kconfig b/drivers/dfu/Kconfig index 1cbec811285..1e1dff33d25 100644 --- a/drivers/dfu/Kconfig +++ b/drivers/dfu/Kconfig @@ -54,5 +54,11 @@ config DFU_SF_PART This option enables the support of "part" and "partubi" target in SPI flash DFU back end. +config DFU_MTD + bool "MTD back end for DFU" + depends on MTD + help + This option enables using DFU to read and write to on any MTD device. + endif endmenu diff --git a/drivers/dfu/Makefile b/drivers/dfu/Makefile index 4164f342acc..ebb119f3982 100644 --- a/drivers/dfu/Makefile +++ b/drivers/dfu/Makefile @@ -5,6 +5,7 @@ obj-$(CONFIG_$(SPL_)DFU) += dfu.o obj-$(CONFIG_$(SPL_)DFU_MMC) += dfu_mmc.o +obj-$(CONFIG_$(SPL_)DFU_MTD) += dfu_mtd.o obj-$(CONFIG_$(SPL_)DFU_NAND) += dfu_nand.o obj-$(CONFIG_$(SPL_)DFU_RAM) += dfu_ram.o obj-$(CONFIG_$(SPL_)DFU_SF) += dfu_sf.o diff --git a/drivers/dfu/dfu.c b/drivers/dfu/dfu.c index 50919fcae9a..4f4a07b790c 100644 --- a/drivers/dfu/dfu.c +++ b/drivers/dfu/dfu.c @@ -462,6 +462,9 @@ static int dfu_fill_entity(struct dfu_entity *dfu, char *s, int alt, if (strcmp(interface, "mmc") == 0) { if (dfu_fill_entity_mmc(dfu, devstr, s)) return -1; + } else if (strcmp(interface, "mtd") == 0) { + if (dfu_fill_entity_mtd(dfu, devstr, s)) + return -1; } else if (strcmp(interface, "nand") == 0) { if (dfu_fill_entity_nand(dfu, devstr, s)) return -1; @@ -566,7 +569,7 @@ int dfu_config_entities(char *env, char *interface, char *devstr) const char *dfu_get_dev_type(enum dfu_device_type t) { const char *const dev_t[] = {NULL, "eMMC", "OneNAND", "NAND", "RAM", - "SF"}; + "SF", "MTD"}; return dev_t[t]; } diff --git a/drivers/dfu/dfu_mtd.c b/drivers/dfu/dfu_mtd.c new file mode 100644 index 00000000000..13fc837c69b --- /dev/null +++ b/drivers/dfu/dfu_mtd.c @@ -0,0 +1,232 @@ +// SPDX-License-Identifier: GPL-2.0+ +/* + * dfu_mtd.c -- DFU for MTD device. + * + * Copyright (C) 2019,STMicroelectronics - All Rights Reserved + * + * Based on dfu_nand.c + */ + +#include +#include +#include + +static bool mtd_is_aligned_with_block_size(struct mtd_info *mtd, u64 size) +{ + return !do_div(size, mtd->erasesize); +} + +static int mtd_block_op(enum dfu_op op, struct dfu_entity *dfu, + u64 offset, void *buf, long *len) +{ + u64 off, lim, remaining; + struct mtd_info *mtd = dfu->data.mtd.info; + struct mtd_oob_ops io_op = {}; + int ret = 0; + bool has_pages = mtd->type == MTD_NANDFLASH || + mtd->type == MTD_MLCNANDFLASH; + + /* if buf == NULL return total size of the area */ + if (!buf) { + *len = dfu->data.mtd.size; + return 0; + } + + off = dfu->data.mtd.start + offset + dfu->bad_skip; + lim = dfu->data.mtd.start + dfu->data.mtd.size; + + if (off >= lim) { + printf("Limit reached 0x%llx\n", lim); + *len = 0; + return op == DFU_OP_READ ? 0 : -EIO; + } + /* limit request with the available size */ + if (off + *len >= lim) + *len = lim - off; + + if (!mtd_is_aligned_with_block_size(mtd, off)) { + printf("Offset not aligned with a block (0x%x)\n", + mtd->erasesize); + return 0; + } + + /* first erase */ + if (op == DFU_OP_WRITE) { + struct erase_info erase_op = {}; + + remaining = round_up(*len, mtd->erasesize); + erase_op.mtd = mtd; + erase_op.addr = off; + erase_op.len = mtd->erasesize; + erase_op.scrub = 0; + + while (remaining) { + if (erase_op.addr + remaining > lim) { + printf("Limit reached 0x%llx while erasing at offset 0x%llx\n", + lim, off); + return -EIO; + } + + ret = mtd_erase(mtd, &erase_op); + + if (ret) { + /* Abort if its not a bad block error */ + if (ret != -EIO) { + printf("Failure while erasing at offset 0x%llx\n", + erase_op.fail_addr); + return 0; + } + printf("Skipping bad block at 0x%08llx\n", + erase_op.addr); + } else { + remaining -= mtd->erasesize; + } + + /* Continue erase behind bad block */ + erase_op.addr += mtd->erasesize; + } + } + + io_op.mode = MTD_OPS_AUTO_OOB; + io_op.len = *len; + if (has_pages && io_op.len > mtd->writesize) + io_op.len = mtd->writesize; + io_op.ooblen = 0; + io_op.datbuf = buf; + io_op.oobbuf = NULL; + + /* Loop over to do the actual read/write */ + remaining = *len; + while (remaining) { + if (off + remaining > lim) { + printf("Limit reached 0x%llx while %s at offset 0x%llx\n", + lim, op == DFU_OP_READ ? "reading" : "writing", + off); + if (op == DFU_OP_READ) { + *len -= remaining; + return 0; + } else { + return -EIO; + } + } + + /* Skip the block if it is bad */ + if (mtd_is_aligned_with_block_size(mtd, off) && + mtd_block_isbad(mtd, off)) { + off += mtd->erasesize; + dfu->bad_skip += mtd->erasesize; + continue; + } + + if (op == DFU_OP_READ) + ret = mtd_read_oob(mtd, off, &io_op); + else + ret = mtd_write_oob(mtd, off, &io_op); + + if (ret) { + printf("Failure while %s at offset 0x%llx\n", + op == DFU_OP_READ ? "reading" : "writing", off); + return -EIO; + } + + off += io_op.retlen; + remaining -= io_op.retlen; + io_op.datbuf += io_op.retlen; + io_op.len = remaining; + if (has_pages && io_op.len > mtd->writesize) + io_op.len = mtd->writesize; + } + + return ret; +} + +static int dfu_get_medium_size_mtd(struct dfu_entity *dfu, u64 *size) +{ + *size = dfu->data.mtd.info->size; + + return 0; +} + +static int dfu_read_medium_mtd(struct dfu_entity *dfu, u64 offset, void *buf, + long *len) +{ + int ret = -1; + + switch (dfu->layout) { + case DFU_RAW_ADDR: + ret = mtd_block_op(DFU_OP_READ, dfu, offset, buf, len); + break; + default: + printf("%s: Layout (%s) not (yet) supported!\n", __func__, + dfu_get_layout(dfu->layout)); + } + + return ret; +} + +static int dfu_write_medium_mtd(struct dfu_entity *dfu, + u64 offset, void *buf, long *len) +{ + int ret = -1; + + switch (dfu->layout) { + case DFU_RAW_ADDR: + ret = mtd_block_op(DFU_OP_WRITE, dfu, offset, buf, len); + break; + default: + printf("%s: Layout (%s) not (yet) supported!\n", __func__, + dfu_get_layout(dfu->layout)); + } + + return ret; +} + +static int dfu_flush_medium_mtd(struct dfu_entity *dfu) +{ + return 0; +} + +static unsigned int dfu_polltimeout_mtd(struct dfu_entity *dfu) +{ + return DFU_DEFAULT_POLL_TIMEOUT; +} + +int dfu_fill_entity_mtd(struct dfu_entity *dfu, char *devstr, char *s) +{ + char *st; + struct mtd_info *mtd; + bool has_pages; + + mtd = get_mtd_device_nm(devstr); + if (IS_ERR_OR_NULL(mtd)) + return -ENODEV; + put_mtd_device(mtd); + + dfu->dev_type = DFU_DEV_MTD; + dfu->data.mtd.info = mtd; + + has_pages = mtd->type == MTD_NANDFLASH || mtd->type == MTD_MLCNANDFLASH; + dfu->max_buf_size = has_pages ? mtd->erasesize : 0; + + st = strsep(&s, " "); + if (!strcmp(st, "raw")) { + dfu->layout = DFU_RAW_ADDR; + dfu->data.mtd.start = simple_strtoul(s, &s, 16); + s++; + dfu->data.mtd.size = simple_strtoul(s, &s, 16); + } else { + printf("%s: (%s) not supported!\n", __func__, st); + return -1; + } + + dfu->get_medium_size = dfu_get_medium_size_mtd; + dfu->read_medium = dfu_read_medium_mtd; + dfu->write_medium = dfu_write_medium_mtd; + dfu->flush_medium = dfu_flush_medium_mtd; + dfu->poll_timeout = dfu_polltimeout_mtd; + + /* initial state */ + dfu->inited = 0; + + return 0; +} -- cgit v1.2.3