[U-Boot] [PATCH] debug_print macros support
From: prafulla_wadaskar prafulla@marvell.com
This debug_prints strategy provides- A. pre-formatted debug and print macros to be used in a code B. flexiblility to enable selective debug prints without modifying a source code For more details refer doc/README.debug_prints
Signed-off-by: prafulla_wadaskar prafulla@marvell.com --- config.mk | 6 +++ doc/README.debug_prints | 89 +++++++++++++++++++++++++++++++++++++++++++++++ include/debug_prints.h | 83 +++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 178 insertions(+), 0 deletions(-) create mode 100644 doc/README.debug_prints create mode 100644 include/debug_prints.h
diff --git a/config.mk b/config.mk index b1254e9..6e85cb5 100644 --- a/config.mk +++ b/config.mk @@ -198,6 +198,12 @@ ifeq ($(PCI_CLOCK),PCI_66M) CFLAGS := $(CFLAGS) -DPCI_66M endif
+CFLAGS := $(CFLAGS) $(shell for dbgprn in $(DEBUG_PRINTS); do \ + if [ "$$dbgprn" != "" ]; then \ + echo "-D$$dbgprn "; \ + fi ; \ + done) + #########################################################################
export HPATH HOSTCC HOSTCFLAGS CROSS_COMPILE \ diff --git a/doc/README.debug_prints b/doc/README.debug_prints new file mode 100644 index 0000000..f03cb4c --- /dev/null +++ b/doc/README.debug_prints @@ -0,0 +1,89 @@ +# +# (C) Copyright 2009 +# Marvell Semiconductor <www.marvell.com> +# Prafulla Wadaskar prafulla@marvell.com +# +# See file CREDITS for list of people who contributed to this +# project. +# +# This program is free software; you can redistribute it and/or +# modify it under the terms of the GNU General Public License as +# published by the Free Software Foundation; either version 2 of +# the License, or (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program; if not, write to the Free Software +# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, +# MA 02110-1301 USA +# + +1.1 Debugging Strategy +====================== +This debug strategy provides predefined macros to be used in code + +DEBUGGING LEVELS +0 To disable all debug messages +1 To enable normal debug macro- debug_print +2 To enable flow trace debug macro- debug_print_ftrace +4 To enable interrupt and timer debug macroc- debug_print4 +8 To enable any special debug messages defined by macro- debug_print8 + +1.2 How to use Debugging strategy in a code ? +============================================= + +1. make sure #include <debug_prints.h> line present in c code +2. define following lines at the top in your code + #ifndef <DBG_FLAG_NAME> + #define <DBG_FLAG_NAME> 0 + #endif + #define DEBUG_PRINT <DBG_FLAG_NAME> + DBG_FLAG_NAME is an identification that is used during build to enable + debug prints +3. define DEBUG_PFX to a small string to identify debug message in a code + This is an optional setting, if you don't define DEBUG_PFX, + by default "dbg: " will be used. +4. use the debug_print, debug_ftrace, debug_print4, debug_print8 macros + in a code wjere ever required + +1.3 How to activate debug messages? +==================================== + +Debug messages can be activated during build time by passing desired +debug level + +1. Enabling Debug messages by passing additional parameter to make + This is a recommended method of debug messages implimentation. + this method give flexibility to enable/disable debug messages + during build without modifying code + Additional command line parameters:- + (a) To enable debug_print messages:- + DEBUG_PRINTS=<DBG_FLAG_NAME> + (b) To enable debug_print_ftrace function:- + DEBUG_PRINTS=<DBG_FLAG_NAME>=2 + (c) To enable debug_print4 messages:- + DEBUG_PRINTS=<DBG_FLAG_NAME>=4 + (d) To enable debug_print8 messages:- + DEBUG_PRINTS=<DBG_FLAG_NAME>=8 + (e) you can enable selective debug prints. + for ex. if you want to enable debug_print and debug_print4 messages + then you can pass ORed debug level value (i.e. 1 for debug_print and + 4 for debug_print4 (1 | 4 = 5) + DEBUG_PRINTS=<DBG_FLAG_NAME>=5 + (f) if you want to enable debug_print defined in more than one c files + then you can pass additional debug flag names as + DEBUG_PRINTS="<DBG_FLAG_NAME1>=1 <DBG_FLAG_NAME2>=2 + the above parameters will enable debug_print messages in a code where + <DBG_FLAG_NAME1> is defined and will enable debug_print_ftrace functions + in a code where <DBG_FLAG_NAME2> is defined + +2. Enabling Debug messages by hardcoding in source file + This is simplest implimentation, just define DEBUG_PRINT to + desired debug level and compile the code, the disadvantage of this + method is, it does not offer flexibility and code with debug message + may become part of your release if not taken care properly. + diff --git a/include/debug_prints.h b/include/debug_prints.h new file mode 100644 index 0000000..8a374ff --- /dev/null +++ b/include/debug_prints.h @@ -0,0 +1,83 @@ +/* + * (C) Copyright 2009 + * Marvell Semiconductor <www.marvell.com> + * Prafulla Wadaskar prafulla@marvell.com + * + * See file CREDITS for list of people who contributed to this + * project. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License as + * published by the Free Software Foundation; either version 2 of + * the License, or (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, + * MA 02110-1301 USA + */ + +/* + * refer doc/README.debug_prints for more details + */ + +#ifndef __DEBUG_PRINTS_H__ +#define __DEBUG_PRINTS_H__ + +#ifndef KERN_ERR +#define KERN_ERR "Err :" +#endif + +#ifndef KERN_INFO +#define KERN_INFO "Info:" +#endif + +#ifndef KERN_WARNING +#define KERN_WARNING "Warn:" +#endif + +#ifndef KERN_DBG +#define KERN_DBG "dbg :" +#endif + +#ifndef DEBUG_PRINT_PFX +#define DEBUG_PRINT_PFX "" +#endif + +#ifndef DEBUG_PRINT +#define DEBUG_PRINT 0 +#endif + +#define error_print(format, arg...) \ + printf(KERN_ERR DEBUG_PRINT_PFX format "\n" , ## arg) +#define info_print(format, arg...) \ + printf(KERN_INFO DEBUG_PRINT_PFX format "\n" , ## arg) +#define warn_print(format, arg...) \ + printf(KERN_WARNING DEBUG_PRINT_PFX format "\n" , ## arg) + +#define debug_print(format, arg...) \ + (DEBUG_PRINT & 1) ? \ + (printf(KERN_DBG DEBUG_PRINT_PFX format "\n" , ## arg))\ + : ({do {} while (0);}) + +#define debug_print_ftrace(format, arg...) \ + (DEBUG_PRINT & 2) ? \ + (printf(KERN_DBG DEBUG_PRINT_PFX "%s() called\n", \ + ( __FUNCTION__ ))) : ({do {} while (0);}) + +#define debug_print4(format, arg...) \ + (DEBUG_PRINT & 4) ? \ + (printf(KERN_DBG DEBUG_PRINT_PFX format "\n" , ## arg))\ + : ({do {} while (0);}) + +#define debug_print8(format, arg...) \ + (DEBUG_PRINT & 8) ? \ + (printf(KERN_DBG DEBUG_PRINT_PFX format "\n" , ## arg))\ + : ({do {} while (0);}) + +#endif /* __DEBUG_PRINTS_H__ */
From: prafulla_wadaskar prafulla@marvell.com
In some cases the u-boot.bin need to be processed further to create bootable u-boot binary from boot device This processing may be cpu,soc and/or board spcific bin_dep.sh provides a mechanism to execute bin_dep.sh if present in above platform specific folders
Signed-off-by: prafulla_wadaskar prafulla@marvell.com --- Makefile | 2 + tools/bin_dep.sh | 79 ++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 81 insertions(+), 0 deletions(-) create mode 100755 tools/bin_dep.sh
diff --git a/Makefile b/Makefile index f857641..8bf36ce 100644 --- a/Makefile +++ b/Makefile @@ -318,6 +318,7 @@ $(obj)u-boot.srec: $(obj)u-boot
$(obj)u-boot.bin: $(obj)u-boot $(OBJCOPY) ${OBJCFLAGS} -O binary $< $@ + @./tools/bin_dep.sh $(CPU) $(BOARD) $(VENDOR) $(SOC) $@
$(obj)u-boot.ldr: $(obj)u-boot $(LDR) -T $(CONFIG_BFIN_CPU) -c $@ $< $(LDR_FLAGS) @@ -3506,6 +3507,7 @@ clean: $(obj)tools/gen_eth_addr $(obj)tools/img2srec \ $(obj)tools/mkimage $(obj)tools/mpc86x_clk \ $(obj)tools/ncb $(obj)tools/ubsha1 + @./tools/bin_dep.sh $(CPU) $(BOARD) $(VENDOR) $(SOC) $@ @rm -f $(obj)board/cray/L1/{bootscript.c,bootscript.image} \ $(obj)board/netstar/{eeprom,crcek,crcit,*.srec,*.bin} \ $(obj)board/trab/trab_fkt $(obj)board/voiceblue/eeprom \ diff --git a/tools/bin_dep.sh b/tools/bin_dep.sh new file mode 100755 index 0000000..b32e2ab --- /dev/null +++ b/tools/bin_dep.sh @@ -0,0 +1,79 @@ +#!/bin/sh + +# (C) Copyright 2009 +# Marvell Semiconductor <www.marvell.com> +# Prafulla Wadaskar prafulla@marvell.com +# +# See file CREDITS for list of people who contributed to this +# project. +# +# This program is free software; you can redistribute it and/or +# modify it under the terms of the GNU General Public License as +# published by the Free Software Foundation; either version 2 of +# the License, or (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program; if not, write to the Free Software +# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, +# MA 02110-1301 USA +# + +# Some platform may require additional post processing of u-boot.bin +# to create usefull binaries +# This script will be executed after each make u-boot.bin and make clean operation + +CUR_DIR=`pwd` +CPU_DIR=$CUR_DIR/cpu/$1 +SOC_DIR=$CPU_DIR/$4 +BOARD_DIR=$CUR_DIR/board/$2 +VENDOR_DIR=$CUR_DIR/board/$3 +VENDOR_BOARD_DIR=$CUR_DIR/board/$3/$2 +DEFAULT_ARG=$5 +#default argument will be <binary name> to be processed in case of build +#default argument will be "clean" in case of mrproper/distclean operation +BIN_ARGS=$DEFAULT_ARG + +#echo CPU_DIR $CPU_DIR +#echo SOC_DIR $SOC_DIR +#echo BOARD_DIR $BOARD_DIR +#echo VENDOR_DIR $VENDOR_DIR +#echo VENDOR_BOARD_DIR $VENDOR_BOARD_DIR +#echo BIN_ARGS $BIN_ARGS + +if [ -d $CPU_DIR ]; then + if [ -f $CPU_DIR/bin_dep.sh ]; then + $CPU_DIR/bin_dep.sh $BIN_ARGS $CPU_DIR + else + if [ -d $SOC_DIR ]; then + if [ -f $SOC_DIR/bin_dep.sh ]; then + $SOC_DIR/bin_dep.sh $BIN_ARGS $SOC_DIR + fi + fi + fi +fi + + +if [ -d $BOARD_DIR ]; then + if [ -f $BOARD_DIR/bin_dep.sh ]; then + $BOARD_DIR/bin_dep.sh $BIN_ARGS $BOARD_DIR + fi +else + if [ -d $VENDOR_DIR ]; then + if [ -f $VENDOR_DIR/bin_dep.sh ]; then + $VENDOR_DIR/bin_dep.sh $BIN_ARGS $VENDOR_DIR + else + if [ -d $VENDOR_BOARD_DIR ]; then + if [ -f $VENDOR_BOARD_DIR/bin_dep.sh ]; then + $VENDOR_BOARD_DIR/bin_dep.sh $BIN_ARGS $VENDOR_BOARD_DIR + fi + fi + fi + fi +fi + +
From: prafulla_wadaskar prafulla@marvell.com
Added macronix SF driver for MTD framework MX25L12805D is supported and tested TBD: sector erase implementation, other deivces support
Signed-off-by: prafulla_wadaskar prafulla@marvell.com Tested by: prafulla_wadaskar prafulla@marvell.com --- drivers/mtd/spi/Makefile | 1 + drivers/mtd/spi/macronix.c | 319 +++++++++++++++++++++++++++++++++++++++++++ drivers/mtd/spi/spi_flash.c | 6 +- 3 files changed, 325 insertions(+), 1 deletions(-) create mode 100644 drivers/mtd/spi/macronix.c
diff --git a/drivers/mtd/spi/Makefile b/drivers/mtd/spi/Makefile index 3d4f892..e527eff 100644 --- a/drivers/mtd/spi/Makefile +++ b/drivers/mtd/spi/Makefile @@ -28,6 +28,7 @@ LIB := $(obj)libspi_flash.a COBJS-$(CONFIG_SPI_FLASH) += spi_flash.o COBJS-$(CONFIG_SPI_FLASH_ATMEL) += atmel.o COBJS-$(CONFIG_SPI_FLASH_STMICRO) += stmicro.o +COBJS-$(CONFIG_SPI_FLASH_MACRONIX) += macronix.o
COBJS := $(COBJS-y) SRCS := $(COBJS:.o=.c) diff --git a/drivers/mtd/spi/macronix.c b/drivers/mtd/spi/macronix.c new file mode 100644 index 0000000..f9bce15 --- /dev/null +++ b/drivers/mtd/spi/macronix.c @@ -0,0 +1,319 @@ +/* + * Copyright 2009(C) Marvell International Ltd. and its affiliates + * Prafulla Wadaskar prafulla@marvell.com + * + * Based on drivers/mtd/spi/stmicor.c + * + * Copyright 2008, Network Appliance Inc. + * Jason McMullan mcmullan@netapp.com + * + * Copyright (C) 2004-2007 Freescale Semiconductor, Inc. + * TsiChung Liew (Tsi-Chung.Liew@freescale.com) + * + * See file CREDITS for list of people who contributed to this + * project. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License as + * published by the Free Software Foundation; either version 2 of + * the License, or (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, + * MA 02110-1301 USA + */ +#include <common.h> +#include <malloc.h> +#include <spi_flash.h> + +#include "spi_flash_internal.h" + +/* MX25Pxx-specific commands */ +#define CMD_MX25PXX_WREN 0x06 /* Write Enable */ +#define CMD_MX25PXX_WRDI 0x04 /* Write Disable */ +#define CMD_MX25PXX_RDSR 0x05 /* Read Status Register */ +#define CMD_MX25PXX_WRSR 0x01 /* Write Status Register */ +#define CMD_MX25PXX_READ 0x03 /* Read Data Bytes */ +#define CMD_MX25PXX_FAST_READ 0x0b /* Read Data Bytes at Higher Speed */ +#define CMD_MX25PXX_PP 0x02 /* Page Program */ +#define CMD_MX25PXX_SE 0x20 /* Sector Erase */ +#define CMD_MX25PXX_BE 0xD8 /* Block Erase */ +#define CMD_MX25PXX_CE 0xc7 /* Chip Erase */ +#define CMD_MX25PXX_DP 0xb9 /* Deep Power-down */ +#define CMD_MX25PXX_RES 0xab /* Release from DP, and Read Signature */ + +#define MXIC_ID_MX25P16 0x15 +#define MXIC_ID_MX25P20 0x12 +#define MXIC_ID_MX25P32 0x16 +#define MXIC_ID_MX25P40 0x13 +#define MXIC_ID_MX25P64 0x17 +#define MXIC_ID_MX25P80 0x14 +#define MXIC_ID_MX25P128 0x18 + +#define MACRONIX_SR_WIP (1 << 0) /* Write-in-Progress */ + +struct macronix_spi_flash_params { + u8 idcode1; + u16 page_size; + u16 pages_per_sector; + u16 sectors_per_block; + u16 nr_blocks; + const char *name; +}; + +struct macronix_spi_flash { + const struct macronix_spi_flash_params *params; + struct spi_flash flash; +}; + +static inline struct macronix_spi_flash *to_macronix_spi_flash(struct spi_flash + *flash) +{ + return container_of(flash, struct macronix_spi_flash, flash); +} + +static const struct macronix_spi_flash_params macronix_spi_flash_table[] = { + { + .idcode1 = MXIC_ID_MX25P128, + .page_size = 256, + .pages_per_sector = 16, + .sectors_per_block = 16, + .nr_blocks = 256, + .name = "MX25L12805D", + }, +}; + +static int macronix_wait_ready(struct spi_flash *flash, unsigned long timeout) +{ + struct spi_slave *spi = flash->spi; + unsigned long timebase; + int ret; + u8 status; + u8 cmd[4] = { CMD_MX25PXX_RDSR, 0xff, 0xff, 0xff }; + + ret = spi_xfer(spi, 32, &cmd[0], NULL, SPI_XFER_BEGIN); + if (ret) { + debug("SF: Failed to send command %02x: %d\n", cmd[0], ret); + return ret; + } + + timebase = get_timer(0); + do { + ret = spi_xfer(spi, 8, NULL, &status, 0); + if (ret) + return -1; + + if ((status & MACRONIX_SR_WIP) == 0) + break; + + } while (get_timer(timebase) < timeout); + + spi_xfer(spi, 0, NULL, NULL, SPI_XFER_END); + + if ((status & MACRONIX_SR_WIP) == 0) + return 0; + + /* Timed out */ + return -1; +} + +static int macronix_read_fast(struct spi_flash *flash, + u32 offset, size_t len, void *buf) +{ + struct macronix_spi_flash *stm = to_macronix_spi_flash(flash); + unsigned long page_addr; + unsigned long page_size; + u8 cmd[5]; + + page_size = stm->params->page_size; + page_addr = offset / page_size; + + cmd[0] = CMD_READ_ARRAY_FAST; + cmd[1] = page_addr >> 8; + cmd[2] = page_addr; + cmd[3] = offset % page_size; + cmd[4] = 0x00; + + return spi_flash_read_common(flash, cmd, sizeof(cmd), buf, len); +} + +static int macronix_write(struct spi_flash *flash, + u32 offset, size_t len, const void *buf) +{ + struct macronix_spi_flash *stm = to_macronix_spi_flash(flash); + unsigned long page_addr; + unsigned long byte_addr; + unsigned long page_size; + size_t chunk_len; + size_t actual; + int ret; + u8 cmd[4]; + + page_size = stm->params->page_size; + page_addr = offset / page_size; + byte_addr = offset % page_size; + + ret = spi_claim_bus(flash->spi); + if (ret) { + debug("SF: Unable to claim SPI bus\n"); + return ret; + } + + ret = 0; + for (actual = 0; actual < len; actual += chunk_len) { + chunk_len = min(len - actual, page_size - byte_addr); + + cmd[0] = CMD_MX25PXX_PP; + cmd[1] = page_addr >> 8; + cmd[2] = page_addr; + cmd[3] = byte_addr; + + debug + ("PP: 0x%p => cmd = { 0x%02x 0x%02x%02x%02x } chunk_len = %d\n", + buf + actual, cmd[0], cmd[1], cmd[2], cmd[3], chunk_len); + + ret = spi_flash_cmd(flash->spi, CMD_MX25PXX_WREN, NULL, 0); + if (ret < 0) { + debug("SF: Enabling Write failed\n"); + break; + } + + ret = spi_flash_cmd_write(flash->spi, cmd, 4, + buf + actual, chunk_len); + if (ret < 0) { + debug("SF: STMicro Page Program failed\n"); + break; + } + + ret = macronix_wait_ready(flash, SPI_FLASH_PROG_TIMEOUT); + if (ret < 0) { + debug("SF: STMicro page programming timed out\n"); + break; + } + + page_addr++; + byte_addr = 0; + } + + debug("SF: STMicro: Successfully programmed %u bytes @ 0x%x\n", + len, offset); + + spi_release_bus(flash->spi); + return ret; +} + +int macronix_erase(struct spi_flash *flash, u32 offset, size_t len) +{ + struct macronix_spi_flash *stm = to_macronix_spi_flash(flash); + unsigned long sector_size; + size_t actual; + int ret; + u8 cmd[4]; + + /* + * This function currently uses sector erase only. + * probably speed things up by using bulk erase + * when possible. + */ + + sector_size = stm->params->page_size * stm->params->pages_per_sector + * stm->params->sectors_per_block; + + if (offset % sector_size || len % sector_size) { + debug("SF: Erase offset/length not multiple of sector size\n"); + return -1; + } + + len /= sector_size; + cmd[0] = CMD_MX25PXX_BE; + cmd[2] = 0x00; + cmd[3] = 0x00; + + ret = spi_claim_bus(flash->spi); + if (ret) { + debug("SF: Unable to claim SPI bus\n"); + return ret; + } + + ret = 0; + for (actual = 0; actual < len; actual++) { + cmd[1] = (offset / sector_size) + actual; + + ret = spi_flash_cmd(flash->spi, CMD_MX25PXX_WREN, NULL, 0); + if (ret < 0) { + debug("SF: Enabling Write failed\n"); + break; + } + + ret = spi_flash_cmd_write(flash->spi, cmd, 4, NULL, 0); + if (ret < 0) { + debug("SF: STMicro page erase failed\n"); + break; + } + + /* Up to 2 seconds */ + ret = macronix_wait_ready(flash, 2 * CONFIG_SYS_HZ); + if (ret < 0) { + debug("SF: STMicro page erase timed out\n"); + break; + } + } + + debug("SF: STMicro: Successfully erased %u bytes @ 0x%x\n", + len * sector_size, offset); + + spi_release_bus(flash->spi); + return ret; +} + +struct spi_flash *spi_flash_probe_macronix(struct spi_slave *spi, u8 * idcode) +{ + const struct macronix_spi_flash_params *params; + struct macronix_spi_flash *stm; + unsigned int i; + int ret; + u8 id[3]; + + ret = spi_flash_cmd(spi, CMD_READ_ID, id, sizeof(id)); + if (ret) + return NULL; + + for (i = 0; i < ARRAY_SIZE(macronix_spi_flash_table); i++) { + params = ¯onix_spi_flash_table[i]; + if (params->idcode1 == idcode[2]) { + break; + } + } + + if (i == ARRAY_SIZE(macronix_spi_flash_table)) { + debug("SF: Unsupported STMicro ID %02x\n", id[1]); + return NULL; + } + + stm = malloc(sizeof(struct macronix_spi_flash)); + if (!stm) { + debug("SF: Failed to allocate memory\n"); + return NULL; + } + + stm->params = params; + stm->flash.spi = spi; + stm->flash.name = params->name; + + stm->flash.write = macronix_write; + stm->flash.erase = macronix_erase; + stm->flash.read = macronix_read_fast; + stm->flash.size = params->page_size * params->pages_per_sector + * params->sectors_per_block * params->nr_blocks; + + printf("SF: Detected %s with page size %u, total %u bytes\n", + params->name, params->page_size, stm->flash.size); + + return &stm->flash; +} diff --git a/drivers/mtd/spi/spi_flash.c b/drivers/mtd/spi/spi_flash.c index d1d81af..bafdfe7 100644 --- a/drivers/mtd/spi/spi_flash.c +++ b/drivers/mtd/spi/spi_flash.c @@ -3,7 +3,6 @@ * * Copyright (C) 2008 Atmel Corporation */ -#define DEBUG #include <common.h> #include <malloc.h> #include <spi.h> @@ -134,6 +133,11 @@ struct spi_flash *spi_flash_probe(unsigned int bus, unsigned int cs, flash = spi_flash_probe_atmel(spi, idcode); break; #endif +#ifdef CONFIG_SPI_FLASH_MACRONIX + case 0xc2: + flash = spi_flash_probe_macronix(spi, idcode); + break; +#endif #ifdef CONFIG_SPI_FLASH_STMICRO case 0x20: flash = spi_flash_probe_stmicro(spi, idcode);
On Friday 03 April 2009 07:49:19 Prafulla Wadaskar wrote:
--- a/drivers/mtd/spi/Makefile +++ b/drivers/mtd/spi/Makefile @@ -28,6 +28,7 @@ LIB := $(obj)libspi_flash.a COBJS-$(CONFIG_SPI_FLASH) += spi_flash.o COBJS-$(CONFIG_SPI_FLASH_ATMEL) += atmel.o COBJS-$(CONFIG_SPI_FLASH_STMICRO) += stmicro.o +COBJS-$(CONFIG_SPI_FLASH_MACRONIX) += macronix.o
please keep the list sorted
--- /dev/null +++ b/drivers/mtd/spi/macronix.c @@ -0,0 +1,319 @@
- Based on drivers/mtd/spi/stmicor.c
typo in file name
- You should have received a copy of the GNU General Public License
- along with this program; if not, write to the Free Software
- Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston,
- MA 02110-1301 USA
- */
+#include <common.h>
please add a new line there
+/* MX25Pxx-specific commands */ +#define CMD_MX25PXX_WREN 0x06 /* Write Enable */ +#define CMD_MX25PXX_WRDI 0x04 /* Write Disable */ +#define CMD_MX25PXX_RDSR 0x05 /* Read Status Register */ +#define CMD_MX25PXX_WRSR 0x01 /* Write Status Register */ +#define CMD_MX25PXX_READ 0x03 /* Read Data Bytes */ +#define CMD_MX25PXX_FAST_READ 0x0b /* Read Data Bytes at Higher Speed */ +#define CMD_MX25PXX_PP 0x02 /* Page Program */ +#define CMD_MX25PXX_SE 0x20 /* Sector Erase */ +#define CMD_MX25PXX_BE 0xD8 /* Block Erase */ +#define CMD_MX25PXX_CE 0xc7 /* Chip Erase */ +#define CMD_MX25PXX_DP 0xb9 /* Deep Power-down */ +#define CMD_MX25PXX_RES 0xab /* Release from DP, and Read Signature */
is it really MX25PXX ? or should it be MX25XX ? the P looks like a hold over from the stmicro driver.
+struct macronix_spi_flash {
- const struct macronix_spi_flash_params *params;
- struct spi_flash flash;
+};
the spi_flash struct needs to be the first member in order to prevent crashes when working with the spi flash layer
- u8 cmd[4] = { CMD_MX25PXX_RDSR, 0xff, 0xff, 0xff };
- ret = spi_xfer(spi, 32, &cmd[0], NULL, SPI_XFER_BEGIN);
do you actually need to read/write 4 bytes here ? shouldnt it be: u8 cmd = CMD_MX25PXX_RDSR; ret = spi_xfer(spi, 8, &cmd, NULL, SPI_XFER_BEGIN);
- if (ret) {
debug("SF: Failed to send command %02x: %d\n", cmd[0], ret);
then you'll need to change to "cmd"
debug("SF: STMicro Page Program failed\n");- debug("SF: STMicro: Successfully programmed %u bytes @ 0x%x\n",
len, offset);debug("SF: STMicro page erase failed\n");debug("SF: STMicro page erase timed out\n");- debug("SF: STMicro: Successfully erased %u bytes @ 0x%x\n",
len * sector_size, offset);debug("SF: Unsupported STMicro ID %02x\n", id[1]);
this isnt STMicro anymore ...
/* Up to 2 seconds */ret = macronix_wait_ready(flash, 2 * CONFIG_SYS_HZ);
there's a common flash erase timeout define
+*spi_flash_probe_macronix(struct spi_slave *spi, u8 * idcode)
no space between "*" and "idcode"
- u8 id[3];
- ret = spi_flash_cmd(spi, CMD_READ_ID, id, sizeof(id));
- if (ret)
return NULL;
no need to read id here ... use the idcode passed in to you
if (params->idcode1 == idcode[2]) {break;}
drop the {} braces
--- a/drivers/mtd/spi/spi_flash.c +++ b/drivers/mtd/spi/spi_flash.c @@ -3,7 +3,6 @@
- Copyright (C) 2008 Atmel Corporation
*/ -#define DEBUG #include <common.h> #include <malloc.h> #include <spi.h>
please drop this hunk ... it doesnt belong in this patch (and it's already been merged elsewhere) -mike
-----Original Message----- From: Mike Frysinger [mailto:vapier@gentoo.org] Sent: Friday, April 03, 2009 7:44 PM To: u-boot@lists.denx.de Cc: Prafulla Wadaskar; Ronen Shitrit Subject: Re: [U-Boot] [PATCH] Macronix MX25xx MTD SPI flash driver
On Friday 03 April 2009 07:49:19 Prafulla Wadaskar wrote:
--- a/drivers/mtd/spi/Makefile +++ b/drivers/mtd/spi/Makefile @@ -28,6 +28,7 @@ LIB := $(obj)libspi_flash.a COBJS-$(CONFIG_SPI_FLASH) += spi_flash.o COBJS-$(CONFIG_SPI_FLASH_ATMEL) += atmel.o COBJS-$(CONFIG_SPI_FLASH_STMICRO) += stmicro.o +COBJS-$(CONFIG_SPI_FLASH_MACRONIX) += macronix.o
please keep the list sorted
Done..
--- /dev/null +++ b/drivers/mtd/spi/macronix.c @@ -0,0 +1,319 @@
- Based on drivers/mtd/spi/stmicor.c
typo in file name
Corrected..
- You should have received a copy of the GNU General
Public License
- along with this program; if not, write to the Free Software
- Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston,
- MA 02110-1301 USA
- */
+#include <common.h>
please add a new line there
Added..
+/* MX25Pxx-specific commands */ +#define CMD_MX25PXX_WREN 0x06 /* Write Enable */ +#define CMD_MX25PXX_WRDI 0x04 /* Write Disable */ +#define CMD_MX25PXX_RDSR 0x05 /* Read Status Register */ +#define CMD_MX25PXX_WRSR 0x01 /* Write Status Register */ +#define CMD_MX25PXX_READ 0x03 /* Read Data Bytes */ +#define CMD_MX25PXX_FAST_READ 0x0b /* Read Data
Bytes at Higher Speed */
+#define CMD_MX25PXX_PP 0x02 /* Page Program */ +#define CMD_MX25PXX_SE 0x20 /* Sector Erase */ +#define CMD_MX25PXX_BE 0xD8 /* Block Erase */ +#define CMD_MX25PXX_CE 0xc7 /* Chip Erase */ +#define CMD_MX25PXX_DP 0xb9 /* Deep Power-down */ +#define CMD_MX25PXX_RES 0xab /* Release from
DP, and Read Signature */
is it really MX25PXX ? or should it be MX25XX ? the P looks like a hold over from the stmicro driver.
You are right It should be MX25LXX for 3.0 volt operation And it should be MX27VXX for 2.5 volt operation To address a common driver for both, I am making it MX25XX
+struct macronix_spi_flash {
- const struct macronix_spi_flash_params *params;
- struct spi_flash flash;
+};
the spi_flash struct needs to be the first member in order to prevent crashes when working with the spi flash layer
Done..
- u8 cmd[4] = { CMD_MX25PXX_RDSR, 0xff, 0xff, 0xff };
- ret = spi_xfer(spi, 32, &cmd[0], NULL, SPI_XFER_BEGIN);
do you actually need to read/write 4 bytes here ? shouldnt it be: u8 cmd = CMD_MX25PXX_RDSR; ret = spi_xfer(spi, 8, &cmd, NULL, SPI_XFER_BEGIN);
Yes, I have checked specs too, we need only 8bit transfer here Done...
- if (ret) {
debug("SF: Failed to send command %02x: %d\n",cmd[0], ret);
then you'll need to change to "cmd"
debug("SF: STMicro Page Program failed\n");- debug("SF: STMicro: Successfully programmed %u bytes @ 0x%x\n",
len, offset);debug("SF: STMicro page erase failed\n");debug("SF: STMicro page erase timed out\n");- debug("SF: STMicro: Successfully erased %u bytes @ 0x%x\n",
len * sector_size, offset);debug("SF: Unsupported STMicro ID %02x\n", id[1]);this isnt STMicro anymore ...
Corrected... Very sorry for this :-(
/* Up to 2 seconds */ret = macronix_wait_ready(flash, 2 * CONFIG_SYS_HZ);there's a common flash erase timeout define
Block erase time for Micronix are different than specified in spi_flash_internals.h Hence MICRONIX spefic timouts defined and used in macronix.c
+*spi_flash_probe_macronix(struct spi_slave *spi, u8 * idcode)
no space between "*" and "idcode"
Done..
- u8 id[3];
- ret = spi_flash_cmd(spi, CMD_READ_ID, id, sizeof(id));
- if (ret)
return NULL;no need to read id here ... use the idcode passed in to you
Done..
if (params->idcode1 == idcode[2]) {break;}drop the {} braces
Dropped..
--- a/drivers/mtd/spi/spi_flash.c +++ b/drivers/mtd/spi/spi_flash.c @@ -3,7 +3,6 @@
- Copyright (C) 2008 Atmel Corporation
*/ -#define DEBUG #include <common.h> #include <malloc.h> #include <spi.h>
please drop this hunk ... it doesnt belong in this patch (and it's already been merged elsewhere) -mike
Okay Dropped
I am ready with the updated patch, How should I release it the community? 1. new patch with review feedback (with same name) Or 2. Delta on the top of previous patch
Any suggestion...?
Regards.. Prafulla . .
On Monday 06 April 2009 03:23:30 Prafulla Wadaskar wrote:
/* Up to 2 seconds */ret = macronix_wait_ready(flash, 2 * CONFIG_SYS_HZ);there's a common flash erase timeout define
Block erase time for Micronix are different than specified in spi_flash_internals.h Hence MICRONIX spefic timouts defined and used in macronix.c
every flash will have a different erase time, but i have a hard time believing you picked 2*HZ (2 seconds) because of your part's requirements. more likely you copied that from the stmicro driver (like every other spi flash driver until i cleaned them up). the latest tree increases the common timeouts to seconds rather than milliseconds, so unless your flash is truly truly awful and requires longer than that, you should use the common define.
I am ready with the updated patch, How should I release it the community?
- new patch with review feedback (with same name)
the current accepted practice is for you to: - update the patch with all the comments - add a "v2" to the subject (so it'd be "[PATCH v2] Macronix .....") - in the comments section of the formatted patch (after the --- marker and above the diffstat), add notes about what changed in the updated patch - send it with the reply to header set to the last relevant message in the patches' thread
- Delta on the top of previous patch
nooooooooo dont ever do this :) -mike
-----Original Message----- From: Mike Frysinger [mailto:vapier@gentoo.org] Sent: Monday, April 06, 2009 1:09 PM To: Prafulla Wadaskar Cc: u-boot@lists.denx.de; Ronen Shitrit Subject: Re: [U-Boot] [PATCH] Macronix MX25xx MTD SPI flash driver
On Monday 06 April 2009 03:23:30 Prafulla Wadaskar wrote:
/* Up to 2 seconds */ret = macronix_wait_ready(flash, 2 *CONFIG_SYS_HZ);
there's a common flash erase timeout define
Block erase time for Micronix are different than specified in spi_flash_internals.h Hence MICRONIX spefic timouts defined
and used
in macronix.c
every flash will have a different erase time, but i have a hard time believing you picked 2*HZ (2 seconds) because of your part's requirements. more likely you copied that from the stmicro driver (like every other spi flash driver until i cleaned them up). the latest tree increases the common timeouts to seconds rather than milliseconds, so unless your flash is truly truly awful and requires longer than that, you should use the common define.
The MX25L12805D specifies Block erase max time to 2sec Even sector erase time is higher (300milsec) Specs can be found at (page 37) http://www.macronix.com/QuickPlace/hq/PageLibrary4825740B00298A3B.nsf/h_Inde...
I am ready with the updated patch, How should I release it the community?
- new patch with review feedback (with same name)
the current accepted practice is for you to:
- update the patch with all the comments
- add a "v2" to the subject (so it'd be "[PATCH v2] Macronix .....")
- in the comments section of the formatted patch (after the
--- marker and above the diffstat), add notes about what changed in the updated patch
- send it with the reply to header set to the last relevant
message in the patches' thread
I will do the needful accordingly
- Delta on the top of previous patch
nooooooooo dont ever do this :)
Thanks.... Stored in my mind Regards.. Prafulla . .
-mike
On Monday 06 April 2009 04:27:59 Prafulla Wadaskar wrote:
From: Mike Frysinger [mailto:vapier@gentoo.org]
On Monday 06 April 2009 03:23:30 Prafulla Wadaskar wrote:
/* Up to 2 seconds */ret = macronix_wait_ready(flash, 2 * CONFIG_SYS_HZ);there's a common flash erase timeout define
Block erase time for Micronix are different than specified in spi_flash_internals.h Hence MICRONIX spefic timouts defined and used in macronix.c
every flash will have a different erase time, but i have a hard time believing you picked 2*HZ (2 seconds) because of your part's requirements. more likely you copied that from the stmicro driver (like every other spi flash driver until i cleaned them up). the latest tree increases the common timeouts to seconds rather than milliseconds, so unless your flash is truly truly awful and requires longer than that, you should use the common define.
The MX25L12805D specifies Block erase max time to 2sec
touche, i stand corrected
Even sector erase time is higher (300milsec)
as i said, the latest code is updated to order of seconds. specifically, write is now 2 seconds while erase is 5 seconds. so both of these are OK for your chip and thus you should use them. -mike
From: prafulla_wadaskar prafulla@marvell.com
Added macronix SF driver for MTD framework MX25L12805D is supported and tested TBD: sector erase implementation, other deivces support
Signed-off-by: prafulla_wadaskar prafulla@marvell.com Reviewed by: Mike Frysinger vapier@gentoo.org Tested by: prafulla_wadaskar prafulla@marvell.com
Signed-off-by: prafulla_wadaskar prafulla@marvell.com --- drivers/mtd/spi/Makefile | 1 + drivers/mtd/spi/macronix.c | 312 +++++++++++++++++++++++++++++++++++++++++++ drivers/mtd/spi/spi_flash.c | 5 + 3 files changed, 318 insertions(+), 0 deletions(-) create mode 100644 drivers/mtd/spi/macronix.c
diff --git a/drivers/mtd/spi/Makefile b/drivers/mtd/spi/Makefile index 3d4f892..2332249 100644 --- a/drivers/mtd/spi/Makefile +++ b/drivers/mtd/spi/Makefile @@ -27,6 +27,7 @@ LIB := $(obj)libspi_flash.a
COBJS-$(CONFIG_SPI_FLASH) += spi_flash.o COBJS-$(CONFIG_SPI_FLASH_ATMEL) += atmel.o +COBJS-$(CONFIG_SPI_FLASH_MACRONIX) += macronix.o COBJS-$(CONFIG_SPI_FLASH_STMICRO) += stmicro.o
COBJS := $(COBJS-y) diff --git a/drivers/mtd/spi/macronix.c b/drivers/mtd/spi/macronix.c new file mode 100644 index 0000000..c98d0e4 --- /dev/null +++ b/drivers/mtd/spi/macronix.c @@ -0,0 +1,312 @@ +/* + * Copyright 2009(C) Marvell International Ltd. and its affiliates + * Prafulla Wadaskar prafulla@marvell.com + * + * Based on drivers/mtd/spi/stmicro.c + * + * Copyright 2008, Network Appliance Inc. + * Jason McMullan mcmullan@netapp.com + * + * Copyright (C) 2004-2007 Freescale Semiconductor, Inc. + * TsiChung Liew (Tsi-Chung.Liew@freescale.com) + * + * See file CREDITS for list of people who contributed to this + * project. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License as + * published by the Free Software Foundation; either version 2 of + * the License, or (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, + * MA 02110-1301 USA + */ + +#include <common.h> +#include <malloc.h> +#include <spi_flash.h> + +#include "spi_flash_internal.h" + +/* MX25xx-specific commands */ +#define CMD_MX25XX_WREN 0x06 /* Write Enable */ +#define CMD_MX25XX_WRDI 0x04 /* Write Disable */ +#define CMD_MX25XX_RDSR 0x05 /* Read Status Register */ +#define CMD_MX25XX_WRSR 0x01 /* Write Status Register */ +#define CMD_MX25XX_READ 0x03 /* Read Data Bytes */ +#define CMD_MX25XX_FAST_READ 0x0b /* Read Data Bytes at Higher Speed */ +#define CMD_MX25XX_PP 0x02 /* Page Program */ +#define CMD_MX25XX_SE 0x20 /* Sector Erase */ +#define CMD_MX25XX_BE 0xD8 /* Block Erase */ +#define CMD_MX25XX_CE 0xc7 /* Chip Erase */ +#define CMD_MX25XX_DP 0xb9 /* Deep Power-down */ +#define CMD_MX25XX_RES 0xab /* Release from DP, and Read Signature */ + +#define MXIC_ID_MX2516 0x15 +#define MXIC_ID_MX2520 0x12 +#define MXIC_ID_MX2532 0x16 +#define MXIC_ID_MX2540 0x13 +#define MXIC_ID_MX2564 0x17 +#define MXIC_ID_MX2580 0x14 +#define MXIC_ID_MX25128 0x18 + +#define MACRONIX_SR_WIP (1 << 0) /* Write-in-Progress */ + +struct macronix_spi_flash_params { + u8 idcode1; + u16 page_size; + u16 pages_per_sector; + u16 sectors_per_block; + u16 nr_blocks; + const char *name; +}; + +struct macronix_spi_flash { + struct spi_flash flash; + const struct macronix_spi_flash_params *params; +}; + +static inline struct macronix_spi_flash *to_macronix_spi_flash(struct spi_flash + *flash) +{ + return container_of(flash, struct macronix_spi_flash, flash); +} + +static const struct macronix_spi_flash_params macronix_spi_flash_table[] = { + { + .idcode1 = MXIC_ID_MX25128, + .page_size = 256, + .pages_per_sector = 16, + .sectors_per_block = 16, + .nr_blocks = 256, + .name = "MX25L12805D", + }, +}; + +static int macronix_wait_ready(struct spi_flash *flash, unsigned long timeout) +{ + struct spi_slave *spi = flash->spi; + unsigned long timebase; + int ret; + u8 status; + u8 cmd = CMD_MX25XX_RDSR; + + ret = spi_xfer(spi, 8, &cmd, NULL, SPI_XFER_BEGIN); + if (ret) { + debug("SF: Failed to send command %02x: %d\n", cmd[0], ret); + return ret; + } + + timebase = get_timer(0); + do { + ret = spi_xfer(spi, 8, NULL, &status, 0); + if (ret) + return -1; + + if ((status & MACRONIX_SR_WIP) == 0) + break; + + } while (get_timer(timebase) < timeout); + + spi_xfer(spi, 0, NULL, NULL, SPI_XFER_END); + + if ((status & MACRONIX_SR_WIP) == 0) + return 0; + + /* Timed out */ + return -1; +} + +static int macronix_read_fast(struct spi_flash *flash, + u32 offset, size_t len, void *buf) +{ + struct macronix_spi_flash *stm = to_macronix_spi_flash(flash); + unsigned long page_addr; + unsigned long page_size; + u8 cmd[5]; + + page_size = stm->params->page_size; + page_addr = offset / page_size; + + cmd[0] = CMD_READ_ARRAY_FAST; + cmd[1] = page_addr >> 8; + cmd[2] = page_addr; + cmd[3] = offset % page_size; + cmd[4] = 0x00; + + return spi_flash_read_common(flash, cmd, sizeof(cmd), buf, len); +} + +static int macronix_write(struct spi_flash *flash, + u32 offset, size_t len, const void *buf) +{ + struct macronix_spi_flash *stm = to_macronix_spi_flash(flash); + unsigned long page_addr; + unsigned long byte_addr; + unsigned long page_size; + size_t chunk_len; + size_t actual; + int ret; + u8 cmd[4]; + + page_size = stm->params->page_size; + page_addr = offset / page_size; + byte_addr = offset % page_size; + + ret = spi_claim_bus(flash->spi); + if (ret) { + debug("SF: Unable to claim SPI bus\n"); + return ret; + } + + ret = 0; + for (actual = 0; actual < len; actual += chunk_len) { + chunk_len = min(len - actual, page_size - byte_addr); + + cmd[0] = CMD_MX25XX_PP; + cmd[1] = page_addr >> 8; + cmd[2] = page_addr; + cmd[3] = byte_addr; + + debug + ("PP: 0x%p => cmd = { 0x%02x 0x%02x%02x%02x } chunk_len = %d\n", + buf + actual, cmd[0], cmd[1], cmd[2], cmd[3], chunk_len); + + ret = spi_flash_cmd(flash->spi, CMD_MX25XX_WREN, NULL, 0); + if (ret < 0) { + debug("SF: Enabling Write failed\n"); + break; + } + + ret = spi_flash_cmd_write(flash->spi, cmd, 4, + buf + actual, chunk_len); + if (ret < 0) { + debug("SF: Macronix Page Program failed\n"); + break; + } + + ret = macronix_wait_ready(flash, SPI_FLASH_PROG_TIMEOUT); + if (ret < 0) { + debug("SF: Macronix page programming timed out\n"); + break; + } + + page_addr++; + byte_addr = 0; + } + + debug("SF: Macronix: Successfully programmed %u bytes @ 0x%x\n", + len, offset); + + spi_release_bus(flash->spi); + return ret; +} + +int macronix_erase(struct spi_flash *flash, u32 offset, size_t len) +{ + struct macronix_spi_flash *stm = to_macronix_spi_flash(flash); + unsigned long sector_size; + size_t actual; + int ret; + u8 cmd[4]; + + /* + * This function currently uses sector erase only. + * probably speed things up by using bulk erase + * when possible. + */ + + sector_size = stm->params->page_size * stm->params->pages_per_sector + * stm->params->sectors_per_block; + + if (offset % sector_size || len % sector_size) { + debug("SF: Erase offset/length not multiple of sector size\n"); + return -1; + } + + len /= sector_size; + cmd[0] = CMD_MX25XX_BE; + cmd[2] = 0x00; + cmd[3] = 0x00; + + ret = spi_claim_bus(flash->spi); + if (ret) { + debug("SF: Unable to claim SPI bus\n"); + return ret; + } + + ret = 0; + for (actual = 0; actual < len; actual++) { + cmd[1] = (offset / sector_size) + actual; + + ret = spi_flash_cmd(flash->spi, CMD_MX25XX_WREN, NULL, 0); + if (ret < 0) { + debug("SF: Enabling Write failed\n"); + break; + } + + ret = spi_flash_cmd_write(flash->spi, cmd, 4, NULL, 0); + if (ret < 0) { + debug("SF: Macronix page erase failed\n"); + break; + } + + ret = macronix_wait_ready(flash, SPI_FLASH_PAGE_ERASE_TIMEOUT); + if (ret < 0) { + debug("SF: Macronix page erase timed out\n"); + break; + } + } + + debug("SF: Macronix: Successfully erased %u bytes @ 0x%x\n", + len * sector_size, offset); + + spi_release_bus(flash->spi); + return ret; +} + +struct spi_flash *spi_flash_probe_macronix(struct spi_slave *spi, u8 *idcode) +{ + const struct macronix_spi_flash_params *params; + struct macronix_spi_flash *stm; + unsigned int i; + + for (i = 0; i < ARRAY_SIZE(macronix_spi_flash_table); i++) { + params = ¯onix_spi_flash_table[i]; + if (params->idcode1 == idcode[2]) + break; + } + + if (i == ARRAY_SIZE(macronix_spi_flash_table)) { + debug("SF: Unsupported Macronix ID %02x\n", idcode[1]); + return NULL; + } + + stm = malloc(sizeof(struct macronix_spi_flash)); + if (!stm) { + debug("SF: Failed to allocate memory\n"); + return NULL; + } + + stm->params = params; + stm->flash.spi = spi; + stm->flash.name = params->name; + + stm->flash.write = macronix_write; + stm->flash.erase = macronix_erase; + stm->flash.read = macronix_read_fast; + stm->flash.size = params->page_size * params->pages_per_sector + * params->sectors_per_block * params->nr_blocks; + + printf("SF: Detected %s with page size %u, total %u bytes\n", + params->name, params->page_size, stm->flash.size); + + return &stm->flash; +} diff --git a/drivers/mtd/spi/spi_flash.c b/drivers/mtd/spi/spi_flash.c index d1d81af..be4c3ed 100644 --- a/drivers/mtd/spi/spi_flash.c +++ b/drivers/mtd/spi/spi_flash.c @@ -134,6 +134,11 @@ struct spi_flash *spi_flash_probe(unsigned int bus, unsigned int cs, flash = spi_flash_probe_atmel(spi, idcode); break; #endif +#ifdef CONFIG_SPI_FLASH_MACRONIX + case 0xc2: + flash = spi_flash_probe_macronix(spi, idcode); + break; +#endif #ifdef CONFIG_SPI_FLASH_STMICRO case 0x20: flash = spi_flash_probe_stmicro(spi, idcode);
Sorry Pls ignore this email...
Regards.. Prafulla . .
-----Original Message----- From: Prafulla Wadaskar [mailto:prafulla@marvell.com] Sent: Monday, April 06, 2009 9:23 PM To: u-boot@lists.denx.de Cc: Ronen Shitrit; Prafulla Wadaskar Subject: [PATCH v2] Macroni ha ha ha ha ha
From: prafulla_wadaskar prafulla@marvell.com
Added macronix SF driver for MTD framework MX25L12805D is supported and tested TBD: sector erase implementation, other deivces support
Signed-off-by: prafulla_wadaskar prafulla@marvell.com Reviewed by: Mike Frysinger vapier@gentoo.org Tested by: prafulla_wadaskar prafulla@marvell.com
Signed-off-by: prafulla_wadaskar prafulla@marvell.com
drivers/mtd/spi/Makefile | 1 + drivers/mtd/spi/macronix.c | 312 +++++++++++++++++++++++++++++++++++++++++++ drivers/mtd/spi/spi_flash.c | 5 + 3 files changed, 318 insertions(+), 0 deletions(-) create mode 100644 drivers/mtd/spi/macronix.c
diff --git a/drivers/mtd/spi/Makefile b/drivers/mtd/spi/Makefile index 3d4f892..2332249 100644 --- a/drivers/mtd/spi/Makefile +++ b/drivers/mtd/spi/Makefile @@ -27,6 +27,7 @@ LIB := $(obj)libspi_flash.a
COBJS-$(CONFIG_SPI_FLASH) += spi_flash.o COBJS-$(CONFIG_SPI_FLASH_ATMEL) += atmel.o +COBJS-$(CONFIG_SPI_FLASH_MACRONIX) += macronix.o COBJS-$(CONFIG_SPI_FLASH_STMICRO) += stmicro.o
COBJS := $(COBJS-y) diff --git a/drivers/mtd/spi/macronix.c b/drivers/mtd/spi/macronix.c new file mode 100644 index 0000000..c98d0e4 --- /dev/null +++ b/drivers/mtd/spi/macronix.c @@ -0,0 +1,312 @@ +/*
- Copyright 2009(C) Marvell International Ltd. and its affiliates
- Prafulla Wadaskar prafulla@marvell.com
- Based on drivers/mtd/spi/stmicro.c
- Copyright 2008, Network Appliance Inc.
- Jason McMullan mcmullan@netapp.com
- Copyright (C) 2004-2007 Freescale Semiconductor, Inc.
- TsiChung Liew (Tsi-Chung.Liew@freescale.com)
- See file CREDITS for list of people who contributed to this
- project.
- This program is free software; you can redistribute it and/or
- modify it under the terms of the GNU General Public License as
- published by the Free Software Foundation; either version 2 of
- the License, or (at your option) any later version.
- This program is distributed in the hope that it will be useful,
- but WITHOUT ANY WARRANTY; without even the implied warranty of
- MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
- GNU General Public License for more details.
- You should have received a copy of the GNU General Public License
- along with this program; if not, write to the Free Software
- Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston,
- MA 02110-1301 USA
- */
+#include <common.h> +#include <malloc.h> +#include <spi_flash.h>
+#include "spi_flash_internal.h"
+/* MX25xx-specific commands */ +#define CMD_MX25XX_WREN 0x06 /* Write Enable */ +#define CMD_MX25XX_WRDI 0x04 /* Write Disable */ +#define CMD_MX25XX_RDSR 0x05 /* Read Status Register */ +#define CMD_MX25XX_WRSR 0x01 /* Write Status Register */ +#define CMD_MX25XX_READ 0x03 /* Read Data Bytes */ +#define CMD_MX25XX_FAST_READ 0x0b /* Read Data Bytes at Higher Speed */ +#define CMD_MX25XX_PP 0x02 /* Page Program */ +#define CMD_MX25XX_SE 0x20 /* Sector Erase */ +#define CMD_MX25XX_BE 0xD8 /* Block Erase */ +#define CMD_MX25XX_CE 0xc7 /* Chip Erase */ +#define CMD_MX25XX_DP 0xb9 /* Deep Power-down */ +#define CMD_MX25XX_RES 0xab /* Release from DP, and Read Signature */
+#define MXIC_ID_MX2516 0x15 +#define MXIC_ID_MX2520 0x12 +#define MXIC_ID_MX2532 0x16 +#define MXIC_ID_MX2540 0x13 +#define MXIC_ID_MX2564 0x17 +#define MXIC_ID_MX2580 0x14 +#define MXIC_ID_MX25128 0x18
+#define MACRONIX_SR_WIP (1 << 0) /* Write-in-Progress */
+struct macronix_spi_flash_params {
- u8 idcode1;
- u16 page_size;
- u16 pages_per_sector;
- u16 sectors_per_block;
- u16 nr_blocks;
- const char *name;
+};
+struct macronix_spi_flash {
- struct spi_flash flash;
- const struct macronix_spi_flash_params *params; };
+static inline struct macronix_spi_flash *to_macronix_spi_flash(struct spi_flash
*flash)+{
- return container_of(flash, struct macronix_spi_flash, flash); }
+static const struct macronix_spi_flash_params macronix_spi_flash_table[] = {
- {
.idcode1 = MXIC_ID_MX25128,.page_size = 256,.pages_per_sector = 16,.sectors_per_block = 16,.nr_blocks = 256,.name = "MX25L12805D",- },
+};
+static int macronix_wait_ready(struct spi_flash *flash, unsigned long +timeout) {
- struct spi_slave *spi = flash->spi;
- unsigned long timebase;
- int ret;
- u8 status;
- u8 cmd = CMD_MX25XX_RDSR;
- ret = spi_xfer(spi, 8, &cmd, NULL, SPI_XFER_BEGIN);
- if (ret) {
debug("SF: Failed to send command %02x: %d\n",cmd[0], ret);
return ret;- }
- timebase = get_timer(0);
- do {
ret = spi_xfer(spi, 8, NULL, &status, 0);if (ret)return -1;if ((status & MACRONIX_SR_WIP) == 0)break;- } while (get_timer(timebase) < timeout);
- spi_xfer(spi, 0, NULL, NULL, SPI_XFER_END);
- if ((status & MACRONIX_SR_WIP) == 0)
return 0;- /* Timed out */
- return -1;
+}
+static int macronix_read_fast(struct spi_flash *flash,
u32 offset, size_t len, void *buf) {- struct macronix_spi_flash *stm = to_macronix_spi_flash(flash);
- unsigned long page_addr;
- unsigned long page_size;
- u8 cmd[5];
- page_size = stm->params->page_size;
- page_addr = offset / page_size;
- cmd[0] = CMD_READ_ARRAY_FAST;
- cmd[1] = page_addr >> 8;
- cmd[2] = page_addr;
- cmd[3] = offset % page_size;
- cmd[4] = 0x00;
- return spi_flash_read_common(flash, cmd, sizeof(cmd),
buf, len); }
+static int macronix_write(struct spi_flash *flash,
u32 offset, size_t len, const void *buf) {- struct macronix_spi_flash *stm = to_macronix_spi_flash(flash);
- unsigned long page_addr;
- unsigned long byte_addr;
- unsigned long page_size;
- size_t chunk_len;
- size_t actual;
- int ret;
- u8 cmd[4];
- page_size = stm->params->page_size;
- page_addr = offset / page_size;
- byte_addr = offset % page_size;
- ret = spi_claim_bus(flash->spi);
- if (ret) {
debug("SF: Unable to claim SPI bus\n");return ret;- }
- ret = 0;
- for (actual = 0; actual < len; actual += chunk_len) {
chunk_len = min(len - actual, page_size - byte_addr);cmd[0] = CMD_MX25XX_PP;cmd[1] = page_addr >> 8;cmd[2] = page_addr;cmd[3] = byte_addr;debug("PP: 0x%p => cmd = { 0x%02x 0x%02x%02x%02x} chunk_len = %d\n",
buf + actual, cmd[0], cmd[1], cmd[2],cmd[3], chunk_len);
ret = spi_flash_cmd(flash->spi,CMD_MX25XX_WREN, NULL, 0);
if (ret < 0) {debug("SF: Enabling Write failed\n");break;}ret = spi_flash_cmd_write(flash->spi, cmd, 4,buf + actual, chunk_len);if (ret < 0) {debug("SF: Macronix Page Program failed\n");break;}ret = macronix_wait_ready(flash,SPI_FLASH_PROG_TIMEOUT);
if (ret < 0) {debug("SF: Macronix page programmingtimed out\n");
break;}page_addr++;byte_addr = 0;- }
- debug("SF: Macronix: Successfully programmed %u bytes @ 0x%x\n",
len, offset);- spi_release_bus(flash->spi);
- return ret;
+}
+int macronix_erase(struct spi_flash *flash, u32 offset, size_t len) {
- struct macronix_spi_flash *stm = to_macronix_spi_flash(flash);
- unsigned long sector_size;
- size_t actual;
- int ret;
- u8 cmd[4];
- /*
* This function currently uses sector erase only.* probably speed things up by using bulk erase* when possible.*/- sector_size = stm->params->page_size *
stm->params->pages_per_sector
* stm->params->sectors_per_block;- if (offset % sector_size || len % sector_size) {
debug("SF: Erase offset/length not multiple ofsector size\n");
return -1;- }
- len /= sector_size;
- cmd[0] = CMD_MX25XX_BE;
- cmd[2] = 0x00;
- cmd[3] = 0x00;
- ret = spi_claim_bus(flash->spi);
- if (ret) {
debug("SF: Unable to claim SPI bus\n");return ret;- }
- ret = 0;
- for (actual = 0; actual < len; actual++) {
cmd[1] = (offset / sector_size) + actual;ret = spi_flash_cmd(flash->spi,CMD_MX25XX_WREN, NULL, 0);
if (ret < 0) {debug("SF: Enabling Write failed\n");break;}ret = spi_flash_cmd_write(flash->spi, cmd, 4, NULL, 0);if (ret < 0) {debug("SF: Macronix page erase failed\n");break;}ret = macronix_wait_ready(flash,SPI_FLASH_PAGE_ERASE_TIMEOUT);
if (ret < 0) {debug("SF: Macronix page erase timed out\n");break;}- }
- debug("SF: Macronix: Successfully erased %u bytes @ 0x%x\n",
len * sector_size, offset);- spi_release_bus(flash->spi);
- return ret;
+}
+struct spi_flash *spi_flash_probe_macronix(struct spi_slave *spi, u8 +*idcode) {
- const struct macronix_spi_flash_params *params;
- struct macronix_spi_flash *stm;
- unsigned int i;
- for (i = 0; i < ARRAY_SIZE(macronix_spi_flash_table); i++) {
params = ¯onix_spi_flash_table[i];if (params->idcode1 == idcode[2])break;- }
- if (i == ARRAY_SIZE(macronix_spi_flash_table)) {
debug("SF: Unsupported Macronix ID %02x\n", idcode[1]);return NULL;- }
- stm = malloc(sizeof(struct macronix_spi_flash));
- if (!stm) {
debug("SF: Failed to allocate memory\n");return NULL;- }
- stm->params = params;
- stm->flash.spi = spi;
- stm->flash.name = params->name;
- stm->flash.write = macronix_write;
- stm->flash.erase = macronix_erase;
- stm->flash.read = macronix_read_fast;
- stm->flash.size = params->page_size * params->pages_per_sector
* params->sectors_per_block * params->nr_blocks;- printf("SF: Detected %s with page size %u, total %u bytes\n",
params->name, params->page_size, stm->flash.size);- return &stm->flash;
+} diff --git a/drivers/mtd/spi/spi_flash.c b/drivers/mtd/spi/spi_flash.c index d1d81af..be4c3ed 100644 --- a/drivers/mtd/spi/spi_flash.c +++ b/drivers/mtd/spi/spi_flash.c @@ -134,6 +134,11 @@ struct spi_flash *spi_flash_probe(unsigned int bus, unsigned int cs, flash = spi_flash_probe_atmel(spi, idcode); break; #endif +#ifdef CONFIG_SPI_FLASH_MACRONIX
- case 0xc2:
flash = spi_flash_probe_macronix(spi, idcode);break;+#endif #ifdef CONFIG_SPI_FLASH_STMICRO case 0x20: flash = spi_flash_probe_stmicro(spi, idcode); -- 1.5.3.3
Dear Prafulla Wadaskar,
In message 1239033171-26122-1-git-send-email-prafulla@marvell.com you wrote:
From: prafulla_wadaskar prafulla@marvell.com
Added macronix SF driver for MTD framework MX25L12805D is supported and tested TBD: sector erase implementation, other deivces support
Signed-off-by: prafulla_wadaskar prafulla@marvell.com Reviewed by: Mike Frysinger vapier@gentoo.org Tested by: prafulla_wadaskar prafulla@marvell.com
Do you think a commit messsage "Macroni ha ha ha ha ha" is adequate?
This is not funny. You are only wasting our time.
Wolfgang Denk
I already expressed my apology for the same.. Once again I am very sorry for the same. Clearly my intention was not to waste your time. It was an accident I will take care not to happen such things ahead... Hope you will understand me
Regards.. Prafulla . .
-----Original Message----- From: Wolfgang Denk [mailto:wd@denx.de] Sent: Monday, April 06, 2009 4:32 PM To: Prafulla Wadaskar Cc: u-boot@lists.denx.de; Ronen Shitrit Subject: Re: [U-Boot] [PATCH v2] Macroni ha ha ha ha ha
Dear Prafulla Wadaskar,
In message 1239033171-26122-1-git-send-email-prafulla@marvell.com you wrote:
From: prafulla_wadaskar prafulla@marvell.com
Added macronix SF driver for MTD framework MX25L12805D is supported and tested TBD: sector erase implementation, other deivces support
Signed-off-by: prafulla_wadaskar prafulla@marvell.com
Reviewed by:
Mike Frysinger vapier@gentoo.org Tested by: prafulla_wadaskar prafulla@marvell.com
Do you think a commit messsage "Macroni ha ha ha ha ha" is adequate?
This is not funny. You are only wasting our time.
Wolfgang Denk
-- DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd@denx.de "The two most common things in the universe are hydrogen and stupi- dity."
From: prafulla_wadaskar prafulla@marvell.com
Added macronix SF driver for MTD framework MX25L12805D is supported and tested TBD: sector erase implementation, other deivces support
Change log: v2: typos corrected, MX25PXX renamed as MX25XX, spi_flash struct declared as first member, RDSR command limited to 8bit transfer and cosmetic fixes
Signed-off-by: prafulla_wadaskar prafulla@marvell.com Reviewed by: Mike Frysinger vapier@gentoo.org Tested by: prafulla_wadaskar prafulla@marvell.com
Signed-off-by: prafulla_wadaskar prafulla@marvell.com --- drivers/mtd/spi/Makefile | 1 + drivers/mtd/spi/macronix.c | 312 +++++++++++++++++++++++++++++++++++++++++++ drivers/mtd/spi/spi_flash.c | 5 + 3 files changed, 318 insertions(+), 0 deletions(-) create mode 100644 drivers/mtd/spi/macronix.c
diff --git a/drivers/mtd/spi/Makefile b/drivers/mtd/spi/Makefile index 3d4f892..2332249 100644 --- a/drivers/mtd/spi/Makefile +++ b/drivers/mtd/spi/Makefile @@ -27,6 +27,7 @@ LIB := $(obj)libspi_flash.a
COBJS-$(CONFIG_SPI_FLASH) += spi_flash.o COBJS-$(CONFIG_SPI_FLASH_ATMEL) += atmel.o +COBJS-$(CONFIG_SPI_FLASH_MACRONIX) += macronix.o COBJS-$(CONFIG_SPI_FLASH_STMICRO) += stmicro.o
COBJS := $(COBJS-y) diff --git a/drivers/mtd/spi/macronix.c b/drivers/mtd/spi/macronix.c new file mode 100644 index 0000000..c98d0e4 --- /dev/null +++ b/drivers/mtd/spi/macronix.c @@ -0,0 +1,312 @@ +/* + * Copyright 2009(C) Marvell International Ltd. and its affiliates + * Prafulla Wadaskar prafulla@marvell.com + * + * Based on drivers/mtd/spi/stmicro.c + * + * Copyright 2008, Network Appliance Inc. + * Jason McMullan mcmullan@netapp.com + * + * Copyright (C) 2004-2007 Freescale Semiconductor, Inc. + * TsiChung Liew (Tsi-Chung.Liew@freescale.com) + * + * See file CREDITS for list of people who contributed to this + * project. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License as + * published by the Free Software Foundation; either version 2 of + * the License, or (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, + * MA 02110-1301 USA + */ + +#include <common.h> +#include <malloc.h> +#include <spi_flash.h> + +#include "spi_flash_internal.h" + +/* MX25xx-specific commands */ +#define CMD_MX25XX_WREN 0x06 /* Write Enable */ +#define CMD_MX25XX_WRDI 0x04 /* Write Disable */ +#define CMD_MX25XX_RDSR 0x05 /* Read Status Register */ +#define CMD_MX25XX_WRSR 0x01 /* Write Status Register */ +#define CMD_MX25XX_READ 0x03 /* Read Data Bytes */ +#define CMD_MX25XX_FAST_READ 0x0b /* Read Data Bytes at Higher Speed */ +#define CMD_MX25XX_PP 0x02 /* Page Program */ +#define CMD_MX25XX_SE 0x20 /* Sector Erase */ +#define CMD_MX25XX_BE 0xD8 /* Block Erase */ +#define CMD_MX25XX_CE 0xc7 /* Chip Erase */ +#define CMD_MX25XX_DP 0xb9 /* Deep Power-down */ +#define CMD_MX25XX_RES 0xab /* Release from DP, and Read Signature */ + +#define MXIC_ID_MX2516 0x15 +#define MXIC_ID_MX2520 0x12 +#define MXIC_ID_MX2532 0x16 +#define MXIC_ID_MX2540 0x13 +#define MXIC_ID_MX2564 0x17 +#define MXIC_ID_MX2580 0x14 +#define MXIC_ID_MX25128 0x18 + +#define MACRONIX_SR_WIP (1 << 0) /* Write-in-Progress */ + +struct macronix_spi_flash_params { + u8 idcode1; + u16 page_size; + u16 pages_per_sector; + u16 sectors_per_block; + u16 nr_blocks; + const char *name; +}; + +struct macronix_spi_flash { + struct spi_flash flash; + const struct macronix_spi_flash_params *params; +}; + +static inline struct macronix_spi_flash *to_macronix_spi_flash(struct spi_flash + *flash) +{ + return container_of(flash, struct macronix_spi_flash, flash); +} + +static const struct macronix_spi_flash_params macronix_spi_flash_table[] = { + { + .idcode1 = MXIC_ID_MX25128, + .page_size = 256, + .pages_per_sector = 16, + .sectors_per_block = 16, + .nr_blocks = 256, + .name = "MX25L12805D", + }, +}; + +static int macronix_wait_ready(struct spi_flash *flash, unsigned long timeout) +{ + struct spi_slave *spi = flash->spi; + unsigned long timebase; + int ret; + u8 status; + u8 cmd = CMD_MX25XX_RDSR; + + ret = spi_xfer(spi, 8, &cmd, NULL, SPI_XFER_BEGIN); + if (ret) { + debug("SF: Failed to send command %02x: %d\n", cmd[0], ret); + return ret; + } + + timebase = get_timer(0); + do { + ret = spi_xfer(spi, 8, NULL, &status, 0); + if (ret) + return -1; + + if ((status & MACRONIX_SR_WIP) == 0) + break; + + } while (get_timer(timebase) < timeout); + + spi_xfer(spi, 0, NULL, NULL, SPI_XFER_END); + + if ((status & MACRONIX_SR_WIP) == 0) + return 0; + + /* Timed out */ + return -1; +} + +static int macronix_read_fast(struct spi_flash *flash, + u32 offset, size_t len, void *buf) +{ + struct macronix_spi_flash *stm = to_macronix_spi_flash(flash); + unsigned long page_addr; + unsigned long page_size; + u8 cmd[5]; + + page_size = stm->params->page_size; + page_addr = offset / page_size; + + cmd[0] = CMD_READ_ARRAY_FAST; + cmd[1] = page_addr >> 8; + cmd[2] = page_addr; + cmd[3] = offset % page_size; + cmd[4] = 0x00; + + return spi_flash_read_common(flash, cmd, sizeof(cmd), buf, len); +} + +static int macronix_write(struct spi_flash *flash, + u32 offset, size_t len, const void *buf) +{ + struct macronix_spi_flash *stm = to_macronix_spi_flash(flash); + unsigned long page_addr; + unsigned long byte_addr; + unsigned long page_size; + size_t chunk_len; + size_t actual; + int ret; + u8 cmd[4]; + + page_size = stm->params->page_size; + page_addr = offset / page_size; + byte_addr = offset % page_size; + + ret = spi_claim_bus(flash->spi); + if (ret) { + debug("SF: Unable to claim SPI bus\n"); + return ret; + } + + ret = 0; + for (actual = 0; actual < len; actual += chunk_len) { + chunk_len = min(len - actual, page_size - byte_addr); + + cmd[0] = CMD_MX25XX_PP; + cmd[1] = page_addr >> 8; + cmd[2] = page_addr; + cmd[3] = byte_addr; + + debug + ("PP: 0x%p => cmd = { 0x%02x 0x%02x%02x%02x } chunk_len = %d\n", + buf + actual, cmd[0], cmd[1], cmd[2], cmd[3], chunk_len); + + ret = spi_flash_cmd(flash->spi, CMD_MX25XX_WREN, NULL, 0); + if (ret < 0) { + debug("SF: Enabling Write failed\n"); + break; + } + + ret = spi_flash_cmd_write(flash->spi, cmd, 4, + buf + actual, chunk_len); + if (ret < 0) { + debug("SF: Macronix Page Program failed\n"); + break; + } + + ret = macronix_wait_ready(flash, SPI_FLASH_PROG_TIMEOUT); + if (ret < 0) { + debug("SF: Macronix page programming timed out\n"); + break; + } + + page_addr++; + byte_addr = 0; + } + + debug("SF: Macronix: Successfully programmed %u bytes @ 0x%x\n", + len, offset); + + spi_release_bus(flash->spi); + return ret; +} + +int macronix_erase(struct spi_flash *flash, u32 offset, size_t len) +{ + struct macronix_spi_flash *stm = to_macronix_spi_flash(flash); + unsigned long sector_size; + size_t actual; + int ret; + u8 cmd[4]; + + /* + * This function currently uses sector erase only. + * probably speed things up by using bulk erase + * when possible. + */ + + sector_size = stm->params->page_size * stm->params->pages_per_sector + * stm->params->sectors_per_block; + + if (offset % sector_size || len % sector_size) { + debug("SF: Erase offset/length not multiple of sector size\n"); + return -1; + } + + len /= sector_size; + cmd[0] = CMD_MX25XX_BE; + cmd[2] = 0x00; + cmd[3] = 0x00; + + ret = spi_claim_bus(flash->spi); + if (ret) { + debug("SF: Unable to claim SPI bus\n"); + return ret; + } + + ret = 0; + for (actual = 0; actual < len; actual++) { + cmd[1] = (offset / sector_size) + actual; + + ret = spi_flash_cmd(flash->spi, CMD_MX25XX_WREN, NULL, 0); + if (ret < 0) { + debug("SF: Enabling Write failed\n"); + break; + } + + ret = spi_flash_cmd_write(flash->spi, cmd, 4, NULL, 0); + if (ret < 0) { + debug("SF: Macronix page erase failed\n"); + break; + } + + ret = macronix_wait_ready(flash, SPI_FLASH_PAGE_ERASE_TIMEOUT); + if (ret < 0) { + debug("SF: Macronix page erase timed out\n"); + break; + } + } + + debug("SF: Macronix: Successfully erased %u bytes @ 0x%x\n", + len * sector_size, offset); + + spi_release_bus(flash->spi); + return ret; +} + +struct spi_flash *spi_flash_probe_macronix(struct spi_slave *spi, u8 *idcode) +{ + const struct macronix_spi_flash_params *params; + struct macronix_spi_flash *stm; + unsigned int i; + + for (i = 0; i < ARRAY_SIZE(macronix_spi_flash_table); i++) { + params = ¯onix_spi_flash_table[i]; + if (params->idcode1 == idcode[2]) + break; + } + + if (i == ARRAY_SIZE(macronix_spi_flash_table)) { + debug("SF: Unsupported Macronix ID %02x\n", idcode[1]); + return NULL; + } + + stm = malloc(sizeof(struct macronix_spi_flash)); + if (!stm) { + debug("SF: Failed to allocate memory\n"); + return NULL; + } + + stm->params = params; + stm->flash.spi = spi; + stm->flash.name = params->name; + + stm->flash.write = macronix_write; + stm->flash.erase = macronix_erase; + stm->flash.read = macronix_read_fast; + stm->flash.size = params->page_size * params->pages_per_sector + * params->sectors_per_block * params->nr_blocks; + + printf("SF: Detected %s with page size %u, total %u bytes\n", + params->name, params->page_size, stm->flash.size); + + return &stm->flash; +} diff --git a/drivers/mtd/spi/spi_flash.c b/drivers/mtd/spi/spi_flash.c index d1d81af..be4c3ed 100644 --- a/drivers/mtd/spi/spi_flash.c +++ b/drivers/mtd/spi/spi_flash.c @@ -134,6 +134,11 @@ struct spi_flash *spi_flash_probe(unsigned int bus, unsigned int cs, flash = spi_flash_probe_atmel(spi, idcode); break; #endif +#ifdef CONFIG_SPI_FLASH_MACRONIX + case 0xc2: + flash = spi_flash_probe_macronix(spi, idcode); + break; +#endif #ifdef CONFIG_SPI_FLASH_STMICRO case 0x20: flash = spi_flash_probe_stmicro(spi, idcode);
Hi Mike I missed subject info (i.e. PATCH v2) is this a problem? Shall I send it again?
Regards.. Prafulla . .
-----Original Message----- From: Prafulla Wadaskar [mailto:prafulla@marvell.com] Sent: Monday, April 06, 2009 9:25 PM To: u-boot@lists.denx.de Cc: Ronen Shitrit; Prafulla Wadaskar Subject: [PATCH] Macronix MX25xx MTD SPI flash driver
From: prafulla_wadaskar prafulla@marvell.com
Added macronix SF driver for MTD framework MX25L12805D is supported and tested TBD: sector erase implementation, other deivces support
Change log: v2: typos corrected, MX25PXX renamed as MX25XX, spi_flash struct declared as first member, RDSR command limited to 8bit transfer and cosmetic fixes
Signed-off-by: prafulla_wadaskar prafulla@marvell.com Reviewed by: Mike Frysinger vapier@gentoo.org Tested by: prafulla_wadaskar prafulla@marvell.com
Signed-off-by: prafulla_wadaskar prafulla@marvell.com
drivers/mtd/spi/Makefile | 1 + drivers/mtd/spi/macronix.c | 312 +++++++++++++++++++++++++++++++++++++++++++ drivers/mtd/spi/spi_flash.c | 5 + 3 files changed, 318 insertions(+), 0 deletions(-) create mode 100644 drivers/mtd/spi/macronix.c
diff --git a/drivers/mtd/spi/Makefile b/drivers/mtd/spi/Makefile index 3d4f892..2332249 100644 --- a/drivers/mtd/spi/Makefile +++ b/drivers/mtd/spi/Makefile @@ -27,6 +27,7 @@ LIB := $(obj)libspi_flash.a
COBJS-$(CONFIG_SPI_FLASH) += spi_flash.o COBJS-$(CONFIG_SPI_FLASH_ATMEL) += atmel.o +COBJS-$(CONFIG_SPI_FLASH_MACRONIX) += macronix.o COBJS-$(CONFIG_SPI_FLASH_STMICRO) += stmicro.o
COBJS := $(COBJS-y) diff --git a/drivers/mtd/spi/macronix.c b/drivers/mtd/spi/macronix.c new file mode 100644 index 0000000..c98d0e4 --- /dev/null +++ b/drivers/mtd/spi/macronix.c @@ -0,0 +1,312 @@ +/*
- Copyright 2009(C) Marvell International Ltd. and its affiliates
- Prafulla Wadaskar prafulla@marvell.com
- Based on drivers/mtd/spi/stmicro.c
- Copyright 2008, Network Appliance Inc.
- Jason McMullan mcmullan@netapp.com
- Copyright (C) 2004-2007 Freescale Semiconductor, Inc.
- TsiChung Liew (Tsi-Chung.Liew@freescale.com)
- See file CREDITS for list of people who contributed to this
- project.
- This program is free software; you can redistribute it and/or
- modify it under the terms of the GNU General Public License as
- published by the Free Software Foundation; either version 2 of
- the License, or (at your option) any later version.
- This program is distributed in the hope that it will be useful,
- but WITHOUT ANY WARRANTY; without even the implied warranty of
- MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
- GNU General Public License for more details.
- You should have received a copy of the GNU General Public License
- along with this program; if not, write to the Free Software
- Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston,
- MA 02110-1301 USA
- */
+#include <common.h> +#include <malloc.h> +#include <spi_flash.h>
+#include "spi_flash_internal.h"
+/* MX25xx-specific commands */ +#define CMD_MX25XX_WREN 0x06 /* Write Enable */ +#define CMD_MX25XX_WRDI 0x04 /* Write Disable */ +#define CMD_MX25XX_RDSR 0x05 /* Read Status Register */ +#define CMD_MX25XX_WRSR 0x01 /* Write Status Register */ +#define CMD_MX25XX_READ 0x03 /* Read Data Bytes */ +#define CMD_MX25XX_FAST_READ 0x0b /* Read Data Bytes at Higher Speed */ +#define CMD_MX25XX_PP 0x02 /* Page Program */ +#define CMD_MX25XX_SE 0x20 /* Sector Erase */ +#define CMD_MX25XX_BE 0xD8 /* Block Erase */ +#define CMD_MX25XX_CE 0xc7 /* Chip Erase */ +#define CMD_MX25XX_DP 0xb9 /* Deep Power-down */ +#define CMD_MX25XX_RES 0xab /* Release from DP, and Read Signature */
+#define MXIC_ID_MX2516 0x15 +#define MXIC_ID_MX2520 0x12 +#define MXIC_ID_MX2532 0x16 +#define MXIC_ID_MX2540 0x13 +#define MXIC_ID_MX2564 0x17 +#define MXIC_ID_MX2580 0x14 +#define MXIC_ID_MX25128 0x18
+#define MACRONIX_SR_WIP (1 << 0) /* Write-in-Progress */
+struct macronix_spi_flash_params {
- u8 idcode1;
- u16 page_size;
- u16 pages_per_sector;
- u16 sectors_per_block;
- u16 nr_blocks;
- const char *name;
+};
+struct macronix_spi_flash {
- struct spi_flash flash;
- const struct macronix_spi_flash_params *params; };
+static inline struct macronix_spi_flash *to_macronix_spi_flash(struct spi_flash
*flash)+{
- return container_of(flash, struct macronix_spi_flash, flash); }
+static const struct macronix_spi_flash_params macronix_spi_flash_table[] = {
- {
.idcode1 = MXIC_ID_MX25128,.page_size = 256,.pages_per_sector = 16,.sectors_per_block = 16,.nr_blocks = 256,.name = "MX25L12805D",- },
+};
+static int macronix_wait_ready(struct spi_flash *flash, unsigned long +timeout) {
- struct spi_slave *spi = flash->spi;
- unsigned long timebase;
- int ret;
- u8 status;
- u8 cmd = CMD_MX25XX_RDSR;
- ret = spi_xfer(spi, 8, &cmd, NULL, SPI_XFER_BEGIN);
- if (ret) {
debug("SF: Failed to send command %02x: %d\n",cmd[0], ret);
return ret;- }
- timebase = get_timer(0);
- do {
ret = spi_xfer(spi, 8, NULL, &status, 0);if (ret)return -1;if ((status & MACRONIX_SR_WIP) == 0)break;- } while (get_timer(timebase) < timeout);
- spi_xfer(spi, 0, NULL, NULL, SPI_XFER_END);
- if ((status & MACRONIX_SR_WIP) == 0)
return 0;- /* Timed out */
- return -1;
+}
+static int macronix_read_fast(struct spi_flash *flash,
u32 offset, size_t len, void *buf) {- struct macronix_spi_flash *stm = to_macronix_spi_flash(flash);
- unsigned long page_addr;
- unsigned long page_size;
- u8 cmd[5];
- page_size = stm->params->page_size;
- page_addr = offset / page_size;
- cmd[0] = CMD_READ_ARRAY_FAST;
- cmd[1] = page_addr >> 8;
- cmd[2] = page_addr;
- cmd[3] = offset % page_size;
- cmd[4] = 0x00;
- return spi_flash_read_common(flash, cmd, sizeof(cmd),
buf, len); }
+static int macronix_write(struct spi_flash *flash,
u32 offset, size_t len, const void *buf) {- struct macronix_spi_flash *stm = to_macronix_spi_flash(flash);
- unsigned long page_addr;
- unsigned long byte_addr;
- unsigned long page_size;
- size_t chunk_len;
- size_t actual;
- int ret;
- u8 cmd[4];
- page_size = stm->params->page_size;
- page_addr = offset / page_size;
- byte_addr = offset % page_size;
- ret = spi_claim_bus(flash->spi);
- if (ret) {
debug("SF: Unable to claim SPI bus\n");return ret;- }
- ret = 0;
- for (actual = 0; actual < len; actual += chunk_len) {
chunk_len = min(len - actual, page_size - byte_addr);cmd[0] = CMD_MX25XX_PP;cmd[1] = page_addr >> 8;cmd[2] = page_addr;cmd[3] = byte_addr;debug("PP: 0x%p => cmd = { 0x%02x 0x%02x%02x%02x} chunk_len = %d\n",
buf + actual, cmd[0], cmd[1], cmd[2],cmd[3], chunk_len);
ret = spi_flash_cmd(flash->spi,CMD_MX25XX_WREN, NULL, 0);
if (ret < 0) {debug("SF: Enabling Write failed\n");break;}ret = spi_flash_cmd_write(flash->spi, cmd, 4,buf + actual, chunk_len);if (ret < 0) {debug("SF: Macronix Page Program failed\n");break;}ret = macronix_wait_ready(flash,SPI_FLASH_PROG_TIMEOUT);
if (ret < 0) {debug("SF: Macronix page programmingtimed out\n");
break;}page_addr++;byte_addr = 0;- }
- debug("SF: Macronix: Successfully programmed %u bytes @ 0x%x\n",
len, offset);- spi_release_bus(flash->spi);
- return ret;
+}
+int macronix_erase(struct spi_flash *flash, u32 offset, size_t len) {
- struct macronix_spi_flash *stm = to_macronix_spi_flash(flash);
- unsigned long sector_size;
- size_t actual;
- int ret;
- u8 cmd[4];
- /*
* This function currently uses sector erase only.* probably speed things up by using bulk erase* when possible.*/- sector_size = stm->params->page_size *
stm->params->pages_per_sector
* stm->params->sectors_per_block;- if (offset % sector_size || len % sector_size) {
debug("SF: Erase offset/length not multiple ofsector size\n");
return -1;- }
- len /= sector_size;
- cmd[0] = CMD_MX25XX_BE;
- cmd[2] = 0x00;
- cmd[3] = 0x00;
- ret = spi_claim_bus(flash->spi);
- if (ret) {
debug("SF: Unable to claim SPI bus\n");return ret;- }
- ret = 0;
- for (actual = 0; actual < len; actual++) {
cmd[1] = (offset / sector_size) + actual;ret = spi_flash_cmd(flash->spi,CMD_MX25XX_WREN, NULL, 0);
if (ret < 0) {debug("SF: Enabling Write failed\n");break;}ret = spi_flash_cmd_write(flash->spi, cmd, 4, NULL, 0);if (ret < 0) {debug("SF: Macronix page erase failed\n");break;}ret = macronix_wait_ready(flash,SPI_FLASH_PAGE_ERASE_TIMEOUT);
if (ret < 0) {debug("SF: Macronix page erase timed out\n");break;}- }
- debug("SF: Macronix: Successfully erased %u bytes @ 0x%x\n",
len * sector_size, offset);- spi_release_bus(flash->spi);
- return ret;
+}
+struct spi_flash *spi_flash_probe_macronix(struct spi_slave *spi, u8 +*idcode) {
- const struct macronix_spi_flash_params *params;
- struct macronix_spi_flash *stm;
- unsigned int i;
- for (i = 0; i < ARRAY_SIZE(macronix_spi_flash_table); i++) {
params = ¯onix_spi_flash_table[i];if (params->idcode1 == idcode[2])break;- }
- if (i == ARRAY_SIZE(macronix_spi_flash_table)) {
debug("SF: Unsupported Macronix ID %02x\n", idcode[1]);return NULL;- }
- stm = malloc(sizeof(struct macronix_spi_flash));
- if (!stm) {
debug("SF: Failed to allocate memory\n");return NULL;- }
- stm->params = params;
- stm->flash.spi = spi;
- stm->flash.name = params->name;
- stm->flash.write = macronix_write;
- stm->flash.erase = macronix_erase;
- stm->flash.read = macronix_read_fast;
- stm->flash.size = params->page_size * params->pages_per_sector
* params->sectors_per_block * params->nr_blocks;- printf("SF: Detected %s with page size %u, total %u bytes\n",
params->name, params->page_size, stm->flash.size);- return &stm->flash;
+} diff --git a/drivers/mtd/spi/spi_flash.c b/drivers/mtd/spi/spi_flash.c index d1d81af..be4c3ed 100644 --- a/drivers/mtd/spi/spi_flash.c +++ b/drivers/mtd/spi/spi_flash.c @@ -134,6 +134,11 @@ struct spi_flash *spi_flash_probe(unsigned int bus, unsigned int cs, flash = spi_flash_probe_atmel(spi, idcode); break; #endif +#ifdef CONFIG_SPI_FLASH_MACRONIX
- case 0xc2:
flash = spi_flash_probe_macronix(spi, idcode);break;+#endif #ifdef CONFIG_SPI_FLASH_STMICRO case 0x20: flash = spi_flash_probe_stmicro(spi, idcode); -- 1.5.3.3
On Monday 06 April 2009 06:46:29 Prafulla Wadaskar wrote:
I missed subject info (i.e. PATCH v2) is this a problem? Shall I send it again?
dont worry about it. the [PATCH ...] contents of the summary are dropped when merging into git. however, there is one other thing:
Subject: [PATCH] Macronix MX25xx MTD SPI flash driver
From: prafulla_wadaskar prafulla@marvell.com
Added macronix SF driver for MTD framework MX25L12805D is supported and tested TBD: sector erase implementation, other deivces support
Change log: v2: typos corrected, MX25PXX renamed as MX25XX, spi_flash struct declared as first member, RDSR command limited to 8bit transfer and cosmetic fixes
Signed-off-by: prafulla_wadaskar prafulla@marvell.com Reviewed by: Mike Frysinger vapier@gentoo.org Tested by: prafulla_wadaskar prafulla@marvell.com
Signed-off-by: prafulla_wadaskar prafulla@marvell.com
drivers/mtd/spi/Makefile | 1 +
the patch change log should go under that "---" rather than above it. the patch change log is to make reviewers' lives easier, not to go into the actual commit message. but dont worry about sending another patch, i'll just fix it up when i merge it. -mike
-----Original Message----- From: Mike Frysinger [mailto:vapier@gentoo.org] Sent: Monday, April 06, 2009 4:35 PM To: u-boot@lists.denx.de Cc: Prafulla Wadaskar; Ronen Shitrit Subject: Re: [U-Boot] [PATCH] Macronix MX25xx MTD SPI flash driver
Change log: v2: typos corrected, MX25PXX renamed as MX25XX, spi_flash struct declared as first member, RDSR command limited to 8bit
transfer and
cosmetic fixes
Signed-off-by: prafulla_wadaskar prafulla@marvell.com
Reviewed by:
Mike Frysinger vapier@gentoo.org Tested by: prafulla_wadaskar prafulla@marvell.com
Signed-off-by: prafulla_wadaskar prafulla@marvell.com
drivers/mtd/spi/Makefile | 1 +
the patch change log should go under that "---" rather than above it. the patch change log is to make reviewers' lives easier, not to go into the actual commit message. but dont worry about sending another patch, i'll just fix it up when i merge it. -mike
Okay Mike, Thanks for your understanding. I will keep quiet on this thread waiting this patch to get mainlined.
Regards.. Prafulla . .
On Monday 06 April 2009 11:54:43 Prafulla Wadaskar wrote:
--- a/drivers/mtd/spi/Makefile +++ b/drivers/mtd/spi/Makefile @@ -27,6 +27,7 @@ LIB := $(obj)libspi_flash.a
COBJS-$(CONFIG_SPI_FLASH) += spi_flash.o COBJS-$(CONFIG_SPI_FLASH_ATMEL) += atmel.o +COBJS-$(CONFIG_SPI_FLASH_MACRONIX) += macronix.o COBJS-$(CONFIG_SPI_FLASH_STMICRO) += stmicro.o
when submitting patches, you should be developing them against mainline, not something older. trying to merge this patch would result in conflicts because this file has changed here.
+static int macronix_wait_ready(struct spi_flash *flash, unsigned long ...
- if (ret) {
debug("SF: Failed to send command %02x: %d\n", cmd[0], ret);
looks like you forgot to change "cmd[0]" to "cmd"
since these are trivial and i'm in a good mood, i'll go ahead and fix up the conflicts and add them to my sf branch. but in the future please make sure to integrate these methodologies. -mike
Dear Mike Frysinger,
In message 200904060711.50104.vapier@gentoo.org you wrote:
since these are trivial and i'm in a good mood, i'll go ahead and fix up the conflicts and add them to my sf branch. but in the future please make sure to integrate these methodologies.
Please don't or undo. The Signed-off-by: entry is not OK either, and we cannot change this.
Best regards,
Wolfgang Denk
On Monday 06 April 2009 14:15:23 Wolfgang Denk wrote:
In message Mike Frysinger wrote:
since these are trivial and i'm in a good mood, i'll go ahead and fix up the conflicts and add them to my sf branch. but in the future please make sure to integrate these methodologies.
Please don't or undo. The Signed-off-by: entry is not OK either, and we cannot change this.
how so ?
Prafulla: could you resubmit then ? the cleaned up patch i merged is here: http://git.denx.de/?p=u-boot/u-boot- blackfin.git;a=commitdiff;h=eddc6059da8dd71cbe84522b281b15b26dd927c7
the things i fixed: - your name - duplicate tags - patch changelog in commit - missing change to spi_flash_internal.h - fixed Makefile conflicts - cmd[0] vs cmd error -mike
Added macronix SF driver for MTD framework MX25L12805D is supported and tested TBD: sector erase implementation, other deivces support
Reviewed-by: Mike Frysinger vapier@gentoo.org Signed-off-by: Prafulla Wadaskar prafulla@marvell.com --- Change log: v2: typos corrected, MX25PXX renamed as MX25XX, spi_flash struct declared as first member, struch ptr name stm changed to mcx RDSR command limited to 8bit transfer and cosmetic fixes
drivers/mtd/spi/Makefile | 1 + drivers/mtd/spi/macronix.c | 312 ++++++++++++++++++++++++++++++++++ drivers/mtd/spi/spi_flash.c | 5 + drivers/mtd/spi/spi_flash_internal.h | 1 + 4 files changed, 319 insertions(+), 0 deletions(-) create mode 100644 drivers/mtd/spi/macronix.c
diff --git a/drivers/mtd/spi/Makefile b/drivers/mtd/spi/Makefile index a71b16e..27dcbff 100644 --- a/drivers/mtd/spi/Makefile +++ b/drivers/mtd/spi/Makefile @@ -27,6 +27,7 @@ LIB := $(obj)libspi_flash.a
COBJS-$(CONFIG_SPI_FLASH) += spi_flash.o COBJS-$(CONFIG_SPI_FLASH_ATMEL) += atmel.o +COBJS-$(CONFIG_SPI_FLASH_MACRONIX) += macronix.o COBJS-$(CONFIG_SPI_FLASH_SPANSION) += spansion.o COBJS-$(CONFIG_SPI_FLASH_SST) += sst.o COBJS-$(CONFIG_SPI_FLASH_STMICRO) += stmicro.o diff --git a/drivers/mtd/spi/macronix.c b/drivers/mtd/spi/macronix.c new file mode 100644 index 0000000..e2323df --- /dev/null +++ b/drivers/mtd/spi/macronix.c @@ -0,0 +1,312 @@ +/* + * Copyright 2009(C) Marvell International Ltd. and its affiliates + * Prafulla Wadaskar prafulla@marvell.com + * + * Based on drivers/mtd/spi/stmicro.c + * + * Copyright 2008, Network Appliance Inc. + * Jason McMullan mcmullan@netapp.com + * + * Copyright (C) 2004-2007 Freescale Semiconductor, Inc. + * TsiChung Liew (Tsi-Chung.Liew@freescale.com) + * + * See file CREDITS for list of people who contributed to this + * project. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License as + * published by the Free Software Foundation; either version 2 of + * the License, or (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, + * MA 02110-1301 USA + */ + +#include <common.h> +#include <malloc.h> +#include <spi_flash.h> + +#include "spi_flash_internal.h" + +/* MX25xx-specific commands */ +#define CMD_MX25XX_WREN 0x06 /* Write Enable */ +#define CMD_MX25XX_WRDI 0x04 /* Write Disable */ +#define CMD_MX25XX_RDSR 0x05 /* Read Status Register */ +#define CMD_MX25XX_WRSR 0x01 /* Write Status Register */ +#define CMD_MX25XX_READ 0x03 /* Read Data Bytes */ +#define CMD_MX25XX_FAST_READ 0x0b /* Read Data Bytes at Higher Speed */ +#define CMD_MX25XX_PP 0x02 /* Page Program */ +#define CMD_MX25XX_SE 0x20 /* Sector Erase */ +#define CMD_MX25XX_BE 0xD8 /* Block Erase */ +#define CMD_MX25XX_CE 0xc7 /* Chip Erase */ +#define CMD_MX25XX_DP 0xb9 /* Deep Power-down */ +#define CMD_MX25XX_RES 0xab /* Release from DP, and Read Signature */ + +#define MXIC_ID_MX2516 0x15 +#define MXIC_ID_MX2520 0x12 +#define MXIC_ID_MX2532 0x16 +#define MXIC_ID_MX2540 0x13 +#define MXIC_ID_MX2564 0x17 +#define MXIC_ID_MX2580 0x14 +#define MXIC_ID_MX25128 0x18 + +#define MACRONIX_SR_WIP (1 << 0) /* Write-in-Progress */ + +struct macronix_spi_flash_params { + u8 idcode1; + u16 page_size; + u16 pages_per_sector; + u16 sectors_per_block; + u16 nr_blocks; + const char *name; +}; + +struct macronix_spi_flash { + struct spi_flash flash; + const struct macronix_spi_flash_params *params; +}; + +static inline struct macronix_spi_flash *to_macronix_spi_flash(struct spi_flash + *flash) +{ + return container_of(flash, struct macronix_spi_flash, flash); +} + +static const struct macronix_spi_flash_params macronix_spi_flash_table[] = { + { + .idcode1 = MXIC_ID_MX25128, + .page_size = 256, + .pages_per_sector = 16, + .sectors_per_block = 16, + .nr_blocks = 256, + .name = "MX25L12805D", + }, +}; + +static int macronix_wait_ready(struct spi_flash *flash, unsigned long timeout) +{ + struct spi_slave *spi = flash->spi; + unsigned long timebase; + int ret; + u8 status; + u8 cmd = CMD_MX25XX_RDSR; + + ret = spi_xfer(spi, 8, &cmd, NULL, SPI_XFER_BEGIN); + if (ret) { + debug("SF: Failed to send command %02x: %d\n", cmd, ret); + return ret; + } + + timebase = get_timer(0); + do { + ret = spi_xfer(spi, 8, NULL, &status, 0); + if (ret) + return -1; + + if ((status & MACRONIX_SR_WIP) == 0) + break; + + } while (get_timer(timebase) < timeout); + + spi_xfer(spi, 0, NULL, NULL, SPI_XFER_END); + + if ((status & MACRONIX_SR_WIP) == 0) + return 0; + + /* Timed out */ + return -1; +} + +static int macronix_read_fast(struct spi_flash *flash, + u32 offset, size_t len, void *buf) +{ + struct macronix_spi_flash *mcx = to_macronix_spi_flash(flash); + unsigned long page_addr; + unsigned long page_size; + u8 cmd[5]; + + page_size = mcx->params->page_size; + page_addr = offset / page_size; + + cmd[0] = CMD_READ_ARRAY_FAST; + cmd[1] = page_addr >> 8; + cmd[2] = page_addr; + cmd[3] = offset % page_size; + cmd[4] = 0x00; + + return spi_flash_read_common(flash, cmd, sizeof(cmd), buf, len); +} + +static int macronix_write(struct spi_flash *flash, + u32 offset, size_t len, const void *buf) +{ + struct macronix_spi_flash *mcx = to_macronix_spi_flash(flash); + unsigned long page_addr; + unsigned long byte_addr; + unsigned long page_size; + size_t chunk_len; + size_t actual; + int ret; + u8 cmd[4]; + + page_size = mcx->params->page_size; + page_addr = offset / page_size; + byte_addr = offset % page_size; + + ret = spi_claim_bus(flash->spi); + if (ret) { + debug("SF: Unable to claim SPI bus\n"); + return ret; + } + + ret = 0; + for (actual = 0; actual < len; actual += chunk_len) { + chunk_len = min(len - actual, page_size - byte_addr); + + cmd[0] = CMD_MX25XX_PP; + cmd[1] = page_addr >> 8; + cmd[2] = page_addr; + cmd[3] = byte_addr; + + debug + ("PP: 0x%p => cmd = { 0x%02x 0x%02x%02x%02x } chunk_len = %d\n", + buf + actual, cmd[0], cmd[1], cmd[2], cmd[3], chunk_len); + + ret = spi_flash_cmd(flash->spi, CMD_MX25XX_WREN, NULL, 0); + if (ret < 0) { + debug("SF: Enabling Write failed\n"); + break; + } + + ret = spi_flash_cmd_write(flash->spi, cmd, 4, + buf + actual, chunk_len); + if (ret < 0) { + debug("SF: Macronix Page Program failed\n"); + break; + } + + ret = macronix_wait_ready(flash, SPI_FLASH_PROG_TIMEOUT); + if (ret < 0) { + debug("SF: Macronix page programming timed out\n"); + break; + } + + page_addr++; + byte_addr = 0; + } + + debug("SF: Macronix: Successfully programmed %u bytes @ 0x%x\n", + len, offset); + + spi_release_bus(flash->spi); + return ret; +} + +int macronix_erase(struct spi_flash *flash, u32 offset, size_t len) +{ + struct macronix_spi_flash *mcx = to_macronix_spi_flash(flash); + unsigned long sector_size; + size_t actual; + int ret; + u8 cmd[4]; + + /* + * This function currently uses sector erase only. + * probably speed things up by using bulk erase + * when possible. + */ + + sector_size = mcx->params->page_size * mcx->params->pages_per_sector + * mcx->params->sectors_per_block; + + if (offset % sector_size || len % sector_size) { + debug("SF: Erase offset/length not multiple of sector size\n"); + return -1; + } + + len /= sector_size; + cmd[0] = CMD_MX25XX_BE; + cmd[2] = 0x00; + cmd[3] = 0x00; + + ret = spi_claim_bus(flash->spi); + if (ret) { + debug("SF: Unable to claim SPI bus\n"); + return ret; + } + + ret = 0; + for (actual = 0; actual < len; actual++) { + cmd[1] = (offset / sector_size) + actual; + + ret = spi_flash_cmd(flash->spi, CMD_MX25XX_WREN, NULL, 0); + if (ret < 0) { + debug("SF: Enabling Write failed\n"); + break; + } + + ret = spi_flash_cmd_write(flash->spi, cmd, 4, NULL, 0); + if (ret < 0) { + debug("SF: Macronix page erase failed\n"); + break; + } + + ret = macronix_wait_ready(flash, SPI_FLASH_PAGE_ERASE_TIMEOUT); + if (ret < 0) { + debug("SF: Macronix page erase timed out\n"); + break; + } + } + + debug("SF: Macronix: Successfully erased %u bytes @ 0x%x\n", + len * sector_size, offset); + + spi_release_bus(flash->spi); + return ret; +} + +struct spi_flash *spi_flash_probe_macronix(struct spi_slave *spi, u8 *idcode) +{ + const struct macronix_spi_flash_params *params; + struct macronix_spi_flash *mcx; + unsigned int i; + + for (i = 0; i < ARRAY_SIZE(macronix_spi_flash_table); i++) { + params = ¯onix_spi_flash_table[i]; + if (params->idcode1 == idcode[2]) + break; + } + + if (i == ARRAY_SIZE(macronix_spi_flash_table)) { + debug("SF: Unsupported Macronix ID %02x\n", idcode[1]); + return NULL; + } + + mcx = malloc(sizeof(struct macronix_spi_flash)); + if (!mcx) { + debug("SF: Failed to allocate memory\n"); + return NULL; + } + + mcx->params = params; + mcx->flash.spi = spi; + mcx->flash.name = params->name; + + mcx->flash.write = macronix_write; + mcx->flash.erase = macronix_erase; + mcx->flash.read = macronix_read_fast; + mcx->flash.size = params->page_size * params->pages_per_sector + * params->sectors_per_block * params->nr_blocks; + + printf("SF: Detected %s with page size %u, total %u bytes\n", + params->name, params->page_size, mcx->flash.size); + + return &mcx->flash; +} diff --git a/drivers/mtd/spi/spi_flash.c b/drivers/mtd/spi/spi_flash.c index 274895a..0c83231 100644 --- a/drivers/mtd/spi/spi_flash.c +++ b/drivers/mtd/spi/spi_flash.c @@ -134,6 +134,11 @@ struct spi_flash *spi_flash_probe(unsigned int bus, unsigned int cs, flash = spi_flash_probe_atmel(spi, idcode); break; #endif +#ifdef CONFIG_SPI_FLASH_MACRONIX + case 0xc2: + flash = spi_flash_probe_macronix(spi, idcode); + break; +#endif #ifdef CONFIG_SPI_FLASH_STMICRO case 0x20: flash = spi_flash_probe_stmicro(spi, idcode); diff --git a/drivers/mtd/spi/spi_flash_internal.h b/drivers/mtd/spi/spi_flash_internal.h index 5d1e395..0612383 100644 --- a/drivers/mtd/spi/spi_flash_internal.h +++ b/drivers/mtd/spi/spi_flash_internal.h @@ -46,5 +46,6 @@ int spi_flash_read_common(struct spi_flash *flash, const u8 *cmd, /* Manufacturer-specific probe functions */ struct spi_flash *spi_flash_probe_spansion(struct spi_slave *spi, u8 *idcode); struct spi_flash *spi_flash_probe_atmel(struct spi_slave *spi, u8 *idcode); +struct spi_flash *spi_flash_probe_macronix(struct spi_slave *spi, u8 *idcode); struct spi_flash *spi_flash_probe_sst(struct spi_slave *spi, u8 *idcode); struct spi_flash *spi_flash_probe_stmicro(struct spi_slave *spi, u8 *idcode);
On Tuesday 07 April 2009 10:17:37 Prafulla Wadaskar wrote:
Added macronix SF driver for MTD framework MX25L12805D is supported and tested TBD: sector erase implementation, other deivces support
ive added this to the blackfin repo in the sf branch, thanks -mike
On Monday 06 April 2009 11:54:43 Prafulla Wadaskar wrote:
From: prafulla_wadaskar prafulla@marvell.com
Added macronix SF driver for MTD framework MX25L12805D is supported and tested TBD: sector erase implementation, other deivces support
Change log: v2: typos corrected, MX25PXX renamed as MX25XX, spi_flash struct declared as first member, RDSR command limited to 8bit transfer and cosmetic fixes
Signed-off-by: prafulla_wadaskar prafulla@marvell.com Reviewed by: Mike Frysinger vapier@gentoo.org Tested by: prafulla_wadaskar prafulla@marvell.com
Signed-off-by: prafulla_wadaskar prafulla@marvell.com
also i noticed these tags ... issues: - duplicate Signed-off-by tags - tag should have your proper name (so "Prafulla Wadaskar") - tags are delimited by dashes, not spaces (so "Reviewed-by:") - Tested-by is pretty much implied by Signed-off-by, so you dont need that if the latter is present ... but that isnt a hard rule, and really only applies to the author of the patch rather than other people who have helped out.
i bet you've just misconfigured your GIT_AUTHOR_* vars or ~/.gitconfig (or however you're setting your name) so that it reads "prafulla_wadaskar" ... -mike
Dear Prafulla Wadaskar,
In message 1239033283-26144-1-git-send-email-prafulla@marvell.com you wrote:
Signed-off-by: prafulla_wadaskar prafulla@marvell.com Reviewed by: Mike Frysinger vapier@gentoo.org Tested by: prafulla_wadaskar prafulla@marvell.com
There are a couple of formal issues here:
Please change "Reviewed by:" into "Reviewed-by:" and "Tested by:" into "Tested-by:" - no, this is not nitpicking, some automated tools rely on the correct spelling.
Also, all these entries require a real name, so please change "prafulla_wadaskar" into "Prafulla Wadaskar".
Thanks.
Best regards,
Wolfgang Denk
Dear Prafulla Wadaskar,
In message 1238759359-6544-2-git-send-email-prafulla@marvell.com you wrote:
From: prafulla_wadaskar prafulla@marvell.com
In some cases the u-boot.bin need to be processed further to create bootable u-boot binary from boot device This processing may be cpu,soc and/or board spcific bin_dep.sh provides a mechanism to execute bin_dep.sh if present in above platform specific folders
Signed-off-by: prafulla_wadaskar prafulla@marvell.com
Makefile | 2 + tools/bin_dep.sh | 79 ++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 81 insertions(+), 0 deletions(-) create mode 100755 tools/bin_dep.sh
I don't see use cases for this. This should probably remain out of tree.
Best regards,
Wolfgang Denk
-----Original Message----- From: Wolfgang Denk [mailto:wd@denx.de] Sent: Friday, April 03, 2009 11:25 PM To: Prafulla Wadaskar Cc: u-boot@lists.denx.de; Ronen Shitrit Subject: Re: [U-Boot] [PATCH] bin_dep.sh Support
Dear Prafulla Wadaskar,
In message 1238759359-6544-2-git-send-email-prafulla@marvell.com you wrote:
From: prafulla_wadaskar prafulla@marvell.com
In some cases the u-boot.bin need to be processed further to create bootable u-boot binary from boot device This processing may
be cpu,soc
and/or board spcific bin_dep.sh provides a mechanism to execute bin_dep.sh if present in above platform specific folders
Signed-off-by: prafulla_wadaskar prafulla@marvell.com
Makefile | 2 + tools/bin_dep.sh | 79
++++++++++++++++++++++++++++++++++++++++++++++++++++++
2 files changed, 81 insertions(+), 0 deletions(-) create
mode 100755
tools/bin_dep.sh
I don't see use cases for this. This should probably remain out of tree.
Dear Wolfgang Denk Relevant use case will be Marvell Kirkwood SOC, this patch will be dependency for the same
Regards.. Prafulla . .
Best regards,
Wolfgang Denk
-- DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd@denx.de I can't understand it. I can't even understand the people who can understand it. - Queen Juliana of the Netherlands.
Dear Prafulla Wadaskar,
In message 73173D32E9439E4ABB5151606C3E19E2508402F4@SC-VEXCH1.marvell.com you wrote:
I don't see use cases for this. This should probably remain out of tree.
Relevant use case will be Marvell Kirkwood SOC, this patch will be dependen cy for the same
It is obvious that you can invent some new stuff (or rather reinvent existing stuff) and then use it for your own boards.
The questions remain if this makes sense, and if this should be added to mainline.
From what I've seen from your image processing code, this looks
largely redundant do already existing features, and I see little sense to add board-specific variants for everybody who wants to define his own private "standard" instead of using existing code.
Best regards,
Wolfgang Denk
Hi
I'm not sure we are on the same page here, the Kirkwood has an internal bootROM which can only boot an image if this image is wrapped by a specific header.
The do_image should take the u-boot bin and wrap it with this header.
How do u suggest we will support this header generation for the U-Boot?
Thanks for your help Ronen Shitrit
-----Original Message----- From: Wolfgang Denk [mailto:wd@denx.de] Sent: Monday, April 06, 2009 11:05 AM To: Prafulla Wadaskar Cc: u-boot@lists.denx.de; Ronen Shitrit Subject: Re: [U-Boot] [PATCH] bin_dep.sh Support
Dear Prafulla Wadaskar,
In message 73173D32E9439E4ABB5151606C3E19E2508402F4@SC-VEXCH1.marvell.com you wrote:
I don't see use cases for this. This should probably remain out of tree.
Relevant use case will be Marvell Kirkwood SOC, this patch will be dependen cy for the same
It is obvious that you can invent some new stuff (or rather reinvent existing stuff) and then use it for your own boards.
The questions remain if this makes sense, and if this should be added to mainline.
From what I've seen from your image processing code, this looks
largely redundant do already existing features, and I see little sense to add board-specific variants for everybody who wants to define his own private "standard" instead of using existing code.
Best regards,
Wolfgang Denk
-- DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd@denx.de War isn't a good life, but it's life. -- Kirk, "A Private Little War", stardate 4211.8
Dear Ronen Shitrit,
In message 309002C0DA137042828828FC53D7A9347E13537FE0@IL-MB01.marvell.com you wrote:
I'm not sure we are on the same page here, the Kirkwood has an internal bootROM which can only boot an image if this image is wrapped by a specific header.
What a stupid (read: broken) hardware design.
There are tons of standards for image formats, and even more commonly used formats I would not dare to call standard; but this processor has to invent yet another one?
The do_image should take the u-boot bin and wrap it with this header.
How do u suggest we will support this header generation for the U-Boot?
Fix the hardware design?
Best regards,
Wolfgang Denk
Dear Ronen Shitrit,
In message 309002C0DA137042828828FC53D7A9347E13537FE0@IL-MB01.marvell.com you wrote:
I'm not sure we are on the same page here, the Kirkwood has an internal bootROM which can only boot an image if this image is wrapped by a specific header.
There are tons of standards for image formats, and even more commonly used formats I would not dare to call standard; but this processor has to invent yet another one?
[Ronen Shitrit] I still think we are not on the same page here. This isn't another image format... Why do we need the bootrom? For example the KW doesn't have HW NAND ECC support :( But we still want to support boot from NAND and ve protected from upto 4 bit ECC. So the CPU wakes up and start running code from the bootROM (embedded ROM in the Kirkwood) this code read the NAND and verify its ECC, then according to the binary image header in the NAND it initialize the DRAM for this specific board and copy the binary image from the NAND to the DRAM and start running from there. I hope now its more clear...
The do_image should take the u-boot bin and wrap it with this header.
How do u suggest we will support this header generation for the U-Boot?
Fix the hardware design? [Ronen Shitrit] probably won't happen...
Best regards,
Wolfgang Denk
-- DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd@denx.de "When anyone says `theoretically,' they really mean `not really.'" - David Parnas
Dear Ronen Shitrit,
In message 309002C0DA137042828828FC53D7A9347E13538016@IL-MB01.marvell.com you wrote:
Dear Ronen Shitrit,
In message 309002C0DA137042828828FC53D7A9347E13537FE0@IL-MB01.marvell.com= you wrote:
I'm not sure we are on the same page here, the Kirkwood has an internal bootROM which can only boot an image if this image is wrapped by a specific header.
There are tons of standards for image formats, and even more commonly used formats I would not dare to call standard; but this processor has to invent yet another one?
Could you please use standard quoting rules? See http://www.netmeister.org/news/learn2quote.html
Aplso, please restrict your line lengh to some 70 characters or so.
Thanks.
[Ronen Shitrit] I still think we are not on the same page here. This isn't another image format...
But yes, it is.
Why do we need the bootrom? For example the KW doesn't have HW NAND ECC support :( But we still want to support boot from NAND and ve protected from upto 4 bi t ECC. So the CPU wakes up and start running code from the bootROM (embedded ROM i n the Kirkwood) this code read the NAND and verify its ECC, then according to the binary image header in the NAND it initialize the DRAM for this spec
You see? "binary image header" Here you say yourself that this is a binary image header, i. e. a prorpietary image format.
Fix the hardware design? [Ronen Shitrit] probably won't happen...
Though so.
Best regards,
Wolfgang Denk
On Monday 06 April 2009 04:24:39 Wolfgang Denk wrote:
In message Ronen Shitrit wrote:
I'm not sure we are on the same page here, the Kirkwood has an internal bootROM which can only boot an image if this image is wrapped by a specific header.
What a stupid (read: broken) hardware design.
There are tons of standards for image formats, and even more commonly used formats I would not dare to call standard; but this processor has to invent yet another one?
true there are plenty of formats out there, but there arent really ones geared towards booting a processor. the ELF format is way too bloated/complex (and yet still not flexible enough) to be usable, and that's about the closest applicable format i'm familiar with. beyond the simple "processor will start execution at address XXXX", there isnt any standardization here.
the Blackfin processor too has a custom format (LDR) for booting code. it's akin to the PHDRs in ELF, plus some additional flag bits to control behavior. but in our world, i wrote external utilities to manage ELF->LDR conversion and so i didnt need a dedicated script in u-boot to do that ugly conversion mojo, just a new Makefile target (u-boot.ldr). -mike
Dear Mike,
in message 200904060503.18864.vapier@gentoo.org you wrote:
the Blackfin processor too has a custom format (LDR) for booting code. it's akin to the PHDRs in ELF, plus some additional flag bits to control behavior. but in our world, i wrote external utilities to manage ELF->LDR conversion and so i didnt need a dedicated script in u-boot to do that ugly conversion mojo, just a new Makefile target (u-boot.ldr).
...which seems to be an acceptable (to both sides) compromise to me.
Best regards,
Wolfgang Denk
On Monday 06 April 2009 05:16:02 Wolfgang Denk wrote:
in message Mike wrote:
the Blackfin processor too has a custom format (LDR) for booting code. it's akin to the PHDRs in ELF, plus some additional flag bits to control behavior. but in our world, i wrote external utilities to manage ELF->LDR conversion and so i didnt need a dedicated script in u-boot to do that ugly conversion mojo, just a new Makefile target (u-boot.ldr).
...which seems to be an acceptable (to both sides) compromise to me.
minus support for embedding the environment into the image ... i have a patch that adds an option to tools/envcrc to export the environment as a binary blob so that it can be passed to the external utility, but you wouldnt take it because you wanted me to fix my hardware instead ;). -mike
Dear Mike Frysinger,
In message 200904060533.47911.vapier@gentoo.org you wrote:
...which seems to be an acceptable (to both sides) compromise to me.
minus support for embedding the environment into the image ... i have a patch that adds an option to tools/envcrc to export the environment as a binary blob so that it can be passed to the external utility, but you wouldnt take it because you wanted me to fix my hardware instead ;).
Not true. I can see that such a tool makes sense - it can be always usefult if you want to pre-initialize the environment on boards where the environment sector(s) is not embedded with the U-Boot image itself.
Just post your code :-)
Best regards,
Wolfgang Denk
On Monday 06 April 2009 06:28:33 Wolfgang Denk wrote:
In message Mike Frysinger wrote:
...which seems to be an acceptable (to both sides) compromise to me.
minus support for embedding the environment into the image ... i have a patch that adds an option to tools/envcrc to export the environment as a binary blob so that it can be passed to the external utility, but you wouldnt take it because you wanted me to fix my hardware instead ;).
Not true. I can see that such a tool makes sense - it can be always usefult if you want to pre-initialize the environment on boards where the environment sector(s) is not embedded with the U-Boot image itself.
Just post your code :-)
ok, extracting the env as a blob isnt problematic, it's the addition of "ENV_IS_EMBEDDED_CUSTOM". the env is embedded in the final image that gets burned into flash, but not by u-boot. the external util takes cares of combining the u-boot ELF and the environment blob.
but i guess i can split the patch into (1) add --binary to tools/envcrc and (2) add ENV_IS_EMBEDDED_CUSTOM. that way when (2) gets shot down again, at least (1) was added so it's one less thing for me to maintain out of tree. -mike
-----Original Message----- From: u-boot-bounces@lists.denx.de [mailto:u-boot-bounces@lists.denx.de] On Behalf Of Wolfgang Denk Sent: Monday, April 06, 2009 2:46 PM To: Mike Frysinger Cc: u-boot@lists.denx.de; Ronen Shitrit Subject: Re: [U-Boot] [PATCH] bin_dep.sh Support
Dear Mike,
in message 200904060503.18864.vapier@gentoo.org you wrote:
the Blackfin processor too has a custom format (LDR) for
booting code.
it's akin to the PHDRs in ELF, plus some additional flag
bits to control behavior.
but in our world, i wrote external utilities to manage ELF->LDR conversion and so i didnt need a dedicated script in u-boot
to do that
ugly conversion mojo, just a new Makefile target (u-boot.ldr).
...which seems to be an acceptable (to both sides) compromise to me.
I understood from the discussion so far that- 1. There are several cases where u-boot.bin does not help alone 2. in some cases there are external tools to be used for the conversion but ideally it is not possible to include each tool in a project
Let's consider specific case - a board having serial console, flash, dram, a processor and supported u-boot.
As a user ideally I expect u-boot.bin generated by u-boot make should go onto flash and board should work. But if this doesn't work then how should we educate the uses and address their questions whereas he might have stumped flashing wrong u-boot.bin to his board...?
This is case with Kirkwood. I don't think it is ONLY processor having this boot sequence architecture, there may be several in pipeline...who knows. Kirkwood is today's need to be addressed.
In my opinion objective of any boot loader is to provide flashable binary.
Bin_dep.sh provides a framework inside u-boot, whether to use it or not is sole decision of board/processor specific code. So why not include it :-)
Regards.. Prafulla . .
On Monday 06 April 2009 05:38:59 Prafulla Wadaskar wrote:
As a user ideally I expect u-boot.bin generated by u-boot make should go onto flash and board should work.
i disagree. u-boot.bin should be the exact flat binary of converting from the u-boot ELF file using objdump.
But if this doesn't work then how should we educate the uses and address their questions whereas he might have stumped flashing wrong u-boot.bin to his board...?
yes. we havent had a problem in the Blackfin world with educating people. if the bootrom is used, then they have to use u-boot.ldr. if they're using an older Blackfin part and the one bootmode that supports direct execution (the part starts running at a hardcoded address), then and only then can they use u-boot.bin. -mike
Dear Prafulla Wadaskar,
In message 73173D32E9439E4ABB5151606C3E19E250840309@SC-VEXCH1.marvell.com you wrote:
...which seems to be an acceptable (to both sides) compromise to me.
I understood from the discussion so far that-
- There are several cases where u-boot.bin does not help alone
- in some cases there are external tools to be used for the conversion but ideally it is not possible to include each tool in a project
The U-Boot source tree includes the source code for the U-Boot boot loader. We do not attempt to include source code for all additional tools that may be needed to build a working U-Boot image - we don;t include GCC source code, nor binutils, nor any other tools that are not specific to U-Boot.
As a user ideally I expect u-boot.bin generated by u-boot make should go onto flash and board should work.
Not really. "u-boot.bin" is supposed to be a raw binary image; this may be directly executable on most processors, but depending on where and how it is stored additional provisions may be necessary. Such tools are out of the scope of the U-Boot project.
But if this doesn't work then how should we educate the uses and address their questions whereas he might have stumped flashing wrong u-boot.bin to his board...?
By providing adequate documentation?
You invent a processor with special requirements, so it seems only natural to me that you educate the users of your processor about the requirements.
I don't think it is ONLY processor having this boot sequence architecture, there may be several in pipeline...who knows.
Try and talk to your hardware design team about the problems such a complicated design is causing on the software side...
Bin_dep.sh provides a framework inside u-boot, whether to use it or not is sole decision of board/processor specific code. So why not include it :-)
Because we do not attempt to include all possible code for any proprietary tool that may be needed to implement all the fancy stuff that might be useful for someone.
Best regards,
Wolfgang Denk
On Monday 06 April 2009 11:38:59 Prafulla Wadaskar wrote:
Bin_dep.sh provides a framework inside u-boot, whether to use it or not is sole decision of board/processor specific code. So why not include it :-)
Since I'm currently working with a Kirkwood platform, I would like to "reactivate" this discussion about support for (external) tools for image creation.
On Kirkwood this image header can include the SDRAM intialization parameters (SDRAM controller registers settings). So this information which is currently *not* included in the Kirkwood U-Boot ports is very important and should be included into the U-Boot ports (from my point of view). Enabling support for such external tools or perhaps even including this tool directly into U-Boot seems to be the way to go for me. Then we could add this additional information like SDRAM init parameters into the U-Boot build process again. And it would be integrated into the mainline U-Boot git repository which is really important from my point of view.
Best regards, Stefan
===================================================================== DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: +49-8142-66989-0 Fax: +49-8142-66989-80 Email: office@denx.de =====================================================================
On Mon, Apr 06, 2009 at 10:24:39AM +0200, Wolfgang Denk wrote:
Dear Ronen Shitrit,
In message 309002C0DA137042828828FC53D7A9347E13537FE0@IL-MB01.marvell.com you wrote:
I'm not sure we are on the same page here, the Kirkwood has an internal bootROM which can only boot an image if this image is wrapped by a specific header.
What a stupid (read: broken) hardware design.
There are tons of standards for image formats, and even more commonly used formats I would not dare to call standard; but this processor has to invent yet another one?
The same could be said about u-boot requiring its own image format (ELF and raw binary images can't be passed a device tree or bd_t, AFAICT).
-Scott
Dear Scott Wood,
In message 20090406193049.GB4439@ld0162-tx32.am.freescale.net you wrote:
There are tons of standards for image formats, and even more commonly used formats I would not dare to call standard; but this processor has to invent yet another one?
The same could be said about u-boot requiring its own image format (ELF and raw binary images can't be passed a device tree or bd_t, AFAICT).
When the U-Boot (or rather PPCBoot at that time) development was starte, there was a very important reason for not using ELF: wasting a full 64 kB for the standard ELF file header was unthinkable on most embedded devices of that time.
Now, on fat systems with ample resource on one side, and on the other side with the infrastructure more or less in place to compose a bootm command from small, separate building blocks, it should be not difficult to add such a feature, too.
On the other hand, ELF images are missing a few key points available in U-Boot images so I would not recommend using plain ELF for any system where reliability or just ease of use are important (which was the other part of the rationale that led to that format).
Best regards,
Wolfgang Denk
On Monday 06 April 2009 15:49:43 Wolfgang Denk wrote:
In message Scott Wood wrote:
There are tons of standards for image formats, and even more commonly used formats I would not dare to call standard; but this processor has to invent yet another one?
The same could be said about u-boot requiring its own image format (ELF and raw binary images can't be passed a device tree or bd_t, AFAICT).
When the U-Boot (or rather PPCBoot at that time) development was starte, there was a very important reason for not using ELF: wasting a full 64 kB for the standard ELF file header was unthinkable on most embedded devices of that time.
Now, on fat systems with ample resource on one side, and on the other side with the infrastructure more or less in place to compose a bootm command from small, separate building blocks, it should be not difficult to add such a feature, too.
On the other hand, ELF images are missing a few key points available in U-Boot images so I would not recommend using plain ELF for any system where reliability or just ease of use are important (which was the other part of the rationale that led to that format).
which are the same reasons processors do not implement support for ELF as a booting source ...
lets say i was to design a brand new part today with an on-chip rom that supports booting from a variety of sources (UART, SPI/I2C/NAND flashes, directly addressable NOR flashes, as a slave device to SPI/I2C, etc...), what format would you recommend that could satisfy all of these requirements ? i'm not aware of any which is why the Blackfin processor has its own stripped down LDR format. -mike
Hi,
Dear Prafulla Wadaskar,
In message 1238759359-6544-2-git-send-email-prafulla@marvell.com you wrote:
From: prafulla_wadaskar prafulla@marvell.com
In some cases the u-boot.bin need to be processed further to create bootable u-boot binary from boot device This processing may be cpu,soc and/or board spcific bin_dep.sh provides a mechanism to execute bin_dep.sh if present in above platform specific folders
Signed-off-by: prafulla_wadaskar prafulla@marvell.com
Makefile | 2 + tools/bin_dep.sh | 79 ++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 81 insertions(+), 0 deletions(-) create mode 100755 tools/bin_dep.sh
I don't see use cases for this. This should probably remain out of tree.
If the wrapping is needed to get a working U-Boot image for this specific board, I'll vote to add the code.
On the other hand, I wouldn't probably add this to the u-boot.bin target. A makefile should really specify what file is generated from what and generating an extra target in the u-boot.bin rule somewhat violates this rule.
I'd say, add another target and the script. Hopefully we can make this as generic as possible so that other boards needing this can only plug in their script.
Cheers Detlev
-----Original Message----- From: Detlev Zundel [mailto:dzu@denx.de] Sent: Monday, April 06, 2009 3:10 PM To: Wolfgang Denk Cc: Prafulla Wadaskar; u-boot@lists.denx.de; Ronen Shitrit Subject: Re: [U-Boot] [PATCH] bin_dep.sh Support
Hi,
Dear Prafulla Wadaskar,
In message
1238759359-6544-2-git-send-email-prafulla@marvell.com you wrote:
From: prafulla_wadaskar prafulla@marvell.com
In some cases the u-boot.bin need to be processed further
to create
bootable u-boot binary from boot device This processing may be cpu,soc and/or board spcific bin_dep.sh provides a mechanism to execute bin_dep.sh if present in above platform specific folders
Signed-off-by: prafulla_wadaskar prafulla@marvell.com
Makefile | 2 + tools/bin_dep.sh | 79
++++++++++++++++++++++++++++++++++++++++++++++++++++++
2 files changed, 81 insertions(+), 0 deletions(-) create mode 100755 tools/bin_dep.sh
I don't see use cases for this. This should probably remain out of tree.
If the wrapping is needed to get a working U-Boot image for this specific board, I'll vote to add the code.
Dear Detlev Thanks for your vote.
Regards.. Prafulla . .
Dear Prafulla Wadaskar,
In message 1238759359-6544-1-git-send-email-prafulla@marvell.com you wrote:
From: prafulla_wadaskar prafulla@marvell.com
This debug_prints strategy provides- A. pre-formatted debug and print macros to be used in a code B. flexiblility to enable selective debug prints without modifying a source code For more details refer doc/README.debug_prints
Signed-off-by: prafulla_wadaskar prafulla@marvell.com
This seems to be a lot of effort, from code to command line pollution - for what exactly?
Which is the problem you are trying to solve?
Best regards,
Wolfgang Denk
Dear Wolfgang Denk
-----Original Message----- From: Wolfgang Denk [mailto:wd@denx.de] Sent: Friday, April 03, 2009 11:23 PM To: Prafulla Wadaskar Cc: u-boot@lists.denx.de; Ronen Shitrit Subject: Re: [U-Boot] [PATCH] debug_print macros support
Dear Prafulla Wadaskar,
In message 1238759359-6544-1-git-send-email-prafulla@marvell.com you wrote:
From: prafulla_wadaskar prafulla@marvell.com
This debug_prints strategy provides- A. pre-formatted debug and print macros to be used in a code B. flexiblility to enable selective debug prints without modifying a source code For more details refer doc/README.debug_prints
Signed-off-by: prafulla_wadaskar prafulla@marvell.com
This seems to be a lot of effort, from code to command line pollution
- for what exactly?
Which is the problem you are trying to solve?
If I understood current u-boot debugging properly, the debug statements are enabled by editing src file by adding "#define DEBUG" Doing this the release contents will be tampered, this may not be relevant from development point of view. But the same release will be used by validation/testing/QA guys and to debug anything out of released contents, no one would like to tamper the source code subject to condition proper debugging strategy is in place.
Secondly, even during development it is a pain to enable/disable debug statements for their needs and finally remove them before release. I suffered this pain that's why discovered this strategy.
The suggested patch provides totally optional strategy that offers- A) enabling selective debug statements without modifying source code B) enabling selective debug statements only from specific source file/s C) Last but not least- simple, small, easy to understand code.
Again.. it is an optional value added tool for u-boot developers :-)
Regards.. Prafulla . .
Best regards,
Wolfgang Denk
-- DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd@denx.de Wisdom is one of the few things that looks bigger the further away it is. - Terry Pratchett, _Witches Abroad_
On Monday 06 April 2009 01:32:30 Prafulla Wadaskar wrote:
If I understood current u-boot debugging properly, the debug statements are enabled by editing src file by adding "#define DEBUG" Doing this the release contents will be tampered, this may not be relevant from development point of view. But the same release will be used by validation/testing/QA guys and to debug anything out of released contents, no one would like to tamper the source code subject to condition proper debugging strategy is in place.
not really true ... you can do: make DBGFLAGS=-DDEBUG
but that will probably enable debug() statements in a lot of things you dont care about ...
i dont really have an opinion on the idea in general -mike
-----Original Message----- From: Mike Frysinger [mailto:vapier@gentoo.org] Sent: Monday, April 06, 2009 12:01 PM To: u-boot@lists.denx.de Cc: Prafulla Wadaskar; Wolfgang Denk; Ronen Shitrit Subject: Re: [U-Boot] [PATCH] debug_print macros support
On Monday 06 April 2009 01:32:30 Prafulla Wadaskar wrote:
If I understood current u-boot debugging properly, the debug statements are enabled by editing src file by adding
"#define DEBUG"
Doing this the release contents will be tampered, this may not be relevant from development point of view. But the same release will be used by validation/testing/QA
guys and to
debug anything out of released contents, no one would like
to tamper
the source code subject to condition proper debugging
strategy is in place.
not really true ... you can do: make DBGFLAGS=-DDEBUG
but that will probably enable debug() statements in a lot of things you dont care about ...
Even this can be resolved by this patch implementation Regards.. Prafulla . .
i dont really have an opinion on the idea in general -mike
Dear Prafulla Wadaskar,
In message 73173D32E9439E4ABB5151606C3E19E2508402D4@SC-VEXCH1.marvell.com you wrote:
This seems to be a lot of effort, from code to command line pollution
- for what exactly?
Which is the problem you are trying to solve?
If I understood current u-boot debugging properly, the debug statements are enabled by editing src file by adding "#define DEBUG" Doing this the release contents will be tampered,
This is just one way to solve this, which is most convenient when you want to enable debugging on a per-file level. Of course you can acchieve the same effect by just passing "-DDEBUG" as a compile option. The Makefile allows to do this globally - there are currently no fine-grained selections available by default.
On the other hand, this is debug code which is not supposed to be present in a production release, so modifying the source seems a minor issue to me.
But the same release will be used by validation/testing/QA guys and to debu g anything out of released contents, no one would like to tamper the source code subject to condition proper debugging strategy is in place.
As mentioned, you can enable this by compile options as well.
The suggested patch provides totally optional strategy that offers- A) enabling selective debug statements without modifying source code B) enabling selective debug statements only from specific source file/s C) Last but not least- simple, small, easy to understand code.
Again.. it is an optional value added tool for u-boot developers :-)
To be honest: I see more overhead added. Looks like overkill to me.
Best regards,
Wolfgang Denk
On Monday 06 April 2009 03:56:39 Wolfgang Denk wrote:
In message Prafulla Wadaskar wrote:
This seems to be a lot of effort, from code to command line pollution
- for what exactly?
Which is the problem you are trying to solve?
If I understood current u-boot debugging properly, the debug statements are enabled by editing src file by adding "#define DEBUG" Doing this the release contents will be tampered,
This is just one way to solve this, which is most convenient when you want to enable debugging on a per-file level. Of course you can acchieve the same effect by just passing "-DDEBUG" as a compile option. The Makefile allows to do this globally - there are currently no fine-grained selections available by default.
i havent tested it, but i think this should get fine-grained flags in a flexible manner ? (the out of tree isnt exact, but you see what i'm after) -mike
diff --git a/config.mk b/config.mk index b1254e9..8791caa 100644 --- a/config.mk +++ b/config.mk @@ -209,20 +209,20 @@ export TEXT_BASE PLATFORM_CPPFLAGS PLATFORM_RELFLAGS CPPFLAGS CFLAGS AFLAGS ifndef REMOTE_BUILD
%.s: %.S - $(CPP) $(AFLAGS) -o $@ $< + $(CPP) $(AFLAGS) $(AFLAGS_$@) -o $@ $< %.o: %.S - $(CC) $(AFLAGS) -c -o $@ $< + $(CC) $(AFLAGS) $(AFLAGS_$@) -c -o $@ $< %.o: %.c - $(CC) $(CFLAGS) -c -o $@ $< + $(CC) $(CFLAGS) $(CFLAGS_$@) -c -o $@ $<
else
$(obj)%.s: %.S - $(CPP) $(AFLAGS) -o $@ $< + $(CPP) $(AFLAGS) $(AFLAGS_$@) -o $@ $< $(obj)%.o: %.S - $(CC) $(AFLAGS) -c -o $@ $< + $(CC) $(AFLAGS) $(AFLAGS_$@) -c -o $@ $< $(obj)%.o: %.c - $(CC) $(CFLAGS) -c -o $@ $< + $(CC) $(CFLAGS) $(CFLAGS_$@) -c -o $@ $< endif
#########################################################################
participants (7)
-
Detlev Zundel -
Mike Frysinger -
Prafulla Wadaskar -
Ronen Shitrit -
Scott Wood -
Stefan Roese -
Wolfgang Denk