April 2009 - U-Boot

[U-Boot] MMC driver for mx31
by alfred steele 15 Apr '09

15 Apr '09

Hi, Is there an available driver for MMC/SD for the mx31 pdk board. If not, what would be the closest candidate for porting it over. Thanks.

1 0

[U-Boot] Uboot md crash on PDK
by alfred steele 14 Apr '09

14 Apr '09

Hi all, I started with trying to tftpload uImage to RAM address 0x10000. It was getting stuck. Hence i thought of doing a memory dump on it. md 0x10000 It too crashed or got stuck. But when i do a md 0x80010000, it works. I checked my linker script to double check on the memory layout. What could have gone wrong when memory dump of specific addresses dont' work. Thanks.

2 6

[U-Boot] Zoom2 respin II
by Tom 14 Apr '09

14 Apr '09

Soon I will be posting a rebase of omap3 zoom2 board with fixes and enhancements per comments. Tom Big changes. GPIO Interface Ported from gpio code from linux. Used in zoom2 debug_board and led code. Add documenation to doc/README.omap3 LED Interface Use existing status_led interface. Add blue led to color led's. Add documentation to doc/README.LED OMAP u-boot.lds Move this from board to cpu. Comments Dirk on 4/4 ... > +static void zoom2_debug_board_detect (void) > +{ > + unsigned int val; > + /* > + * GPIO to query for debug board > + * 158 db board query, bank 5, index 30 > + */ > + gpio_t *gpio5_base = (gpio_t *) OMAP34XX_GPIO5_BASE; > + > + val = __raw_readl (&gpio5_base->datain); Is there any special reason why you use __raw_readl() instead of just readl() here? Looking at ARM's io.h, they seem to map to the same macro? Tom: New gpio interface is used, so this is hidden. BUT looking closely at interface and __raw_readl is still being used. This was kept because I wanted the keep code as close to kernel as possible. ------------------------------------------------------------------------ Dirk 4/4 I'm not sure if I overlooked it in one of the other patches, but if not: Do you like to update doc/README.omap3 Tom: Zoom2 information added to README.omap3 similar to other targets. ------------------------------------------------------------------------ Jean-Christophe Serial 4/2 > COBJS := zoom2.o \ > > - debug_board.o > > + debug_board.o \ > > + zoom2_serial.o it could be nice to disactivate it A : Suprising hard to do. Logic added to do this for LED. COBJS-${CONFIG_STATUS_LED} += led.o Doing something similar leaves the serial_* functions undefined. Defining a set of functions to be weak aliases for these functions does not work cleanly. > +#include "zoom2_serial.h" > > + > > +int zoom2_debug_board_connected (void); please move to a header A : Moved to zoom2_serial.h > + if (zoom2_debug_board_connected ()) { > > + NS16550_t com_port = (NS16550_t) base; > > + int baud_divisor = CONFIG_SYS_NS16550_CLK / 16 / > > + CONFIG_BAUDRATE; aligning it with CONFIG_SYS_NS16550_CLK could be nice please add an empty line A : Done > + com_port->mcr = (MCR_DTR | MCR_RTS); > > + com_port->fcr = (FCR_FIFO_EN | FCR_RXSR | FCR_TXSR); > > + com_port->lcr = LCR_BKSE | LCR_8N1; > > + com_port->dll = baud_divisor & 0xff; > > + com_port->dlm = (baud_divisor >> 8) & 0xff; > > + com_port->lcr = LCR_8N1; clould you explain a fe more what you do here? A : Added a comment that this was generic setup code. > + if (zoom2_debug_board_connected ()) { > > + NS16550_t port = (NS16550_t) base; > > + > > + if (c == '\n') > > + quad_putc_dev (base, '\r'); > > + > > + NS16550_putc (port, c); why not NS16550_putc ((NS16550_t) base, c); A : Done in all the port vs. base functions. > void quad_setbrg_dev (unsigned long base) > > +{ > > + if (zoom2_debug_board_connected ()) { > > + NS16550_t port = (NS16550_t) base; > > + > > + int clock_divisor = CONFIG_SYS_NS16550_CLK / 16 / > > + CONFIG_BAUDRATE; > > + > > + NS16550_reinit (port, clock_divisor); why don't you use the same in in the imit? Tom: No good reason. I was using drivers/serial/serial.c as a guide and this is how it does it, with the clock_divisor function inlined. > + > > +#define QUAD_BASE_0 SERIAL_TL16CP754C_BASE > > +#define QUAD_BASE_1 (SERIAL_TL16CP754C_BASE + 0x100) > > +#define QUAD_BASE_2 (SERIAL_TL16CP754C_BASE + 0x200) > > +#define QUAD_BASE_3 (SERIAL_TL16CP754C_BASE + 0x300) why not #define QUAD_BASE(x) (SERIAL_TL16CP754C_BASE + (x << 8)) Tom: That would work too. I would prefer to keep it as is becuase the register base is less obscured. > +#define QUAD_INIT(n) \ ^^^^^^^^^^^^^^^^^^^^^^^^^^^ white space please fix > > +int quad_init_##n(void) Tom: Done, converted > + enable_gpmc_config(gpmc_config, > > + serial_cs_base, > > + SERIAL_TL16CP754C_BASE, > > + GPMC_SIZE_16M); > > +#endif it's board specific please move to board Tom: Done. > -#ifdef CONFIG_OMAP > > +#if defined(CONFIG_OMAP) && !defined(CONFIG_OMAP3_ZOOM2) a config could be better Tom: I do not use this code, the ifdef is because other omap's do. > #define ONENAND_MAP 0x20000000 /* OneNand addr */ > > /* (actual size small port) */ > > +#define SERIAL_TL16CP754C_BASE 0x10000000 /* Zoom2 Serial chip address */ it's not cpu specfic please move this to the serial driver Tom: Done, moved to zoom2_serial.h > -#define CONFIG_CONS_INDEX 3 > > -#define CONFIG_SYS_NS16550_COM3 OMAP34XX_UART3 > > -#define CONFIG_SERIAL3 3 /* UART3 */ is the removed serial will work on the zoom2? Tom: This, and the other UARTS, are internally connected and should not be used. ------------------------------------------------------------------------ Jean-Christophe Zoom2 base 4/2 From > > + .ARM.extab : { *(.ARM.extab* .gnu.linkonce.armextab.*) } > > + __exidx_start = .; > > + .ARM.exidx : { *(.ARM.exidx* .gnu.linkonce.armexidx.*) } > > + __exidx_end = .; no need please remove > > + > > + . = ALIGN(4); > > + .data : { *(.data) } > > + all the omap3 share the same lds it wil be nice to regroup it as done for the at91 Tom: Done. See OMAP-Consolidate-common-u-boot.lds-to-cpu-layer.patch > +#define CONFIG_CMD_EXT2 /* EXT2 Support */ ^^^^^^^^^^^^^^^^^ white space please fix it > > +#define CONFIG_CMD_FAT /* FAT support */ ditto and the followiing configs too > > +#define CONFIG_CMD_JFFS2 /* JFFS2 Support */ Tom: Done, tab-ified. > +#define CONFIG_JFFS2_PART_OFFSET 0x680000 > > +#define CONFIG_JFFS2_PART_SIZE 0xf980000 /* size of jffs2 */ you do not active the MTD_PARTS it could be usefull Tom: NAND related changes to follow this patch set. > + > > +#define CONFIG_EXTRA_ENV_SETTINGS \ > > + "loadaddr=0x82000000\0" \ load_addr? Tom: This whole section was removed. It came from Zoom1 which came from beagle and does not really apply to zoom2. > + */ > > +#define V_PROMPT "OMAP3 Zoom2# " why? why not only define CONFIG_SYS_PROMPT? Tom: Yes. Good point. Done. > +/* > > + * 2430 has 12 GP timers, they can be driven by the SysClk (12/13/19.2) or by > > + * 32KHz clk, or from external sig. This rate is divided by a local divisor. > > + */ > > +#define V_PVT 7 as said precedently us the 12Mhz as imput with a PVT at 1 will give us a better timer Tom: OK. This code is no longer in the config file. ------------------------------------------------------------------------ Jean-Christophe debug board 4/2 > +COBJS := zoom2.o \ > > + debug_board.o you do not plan to activate it via a CONFIG only? Tom: No. The debug board is ment to be determined at run time. The board is easily detatched and single binary is needed. > + > > +static int debug_board_connected = DEBUG_BOARD_CONNECTED; why not set it as -1; and then avoid the first_time int? Tom: No. I looked at this and the tradeoff would be more code, I would prefer to leave it as it but will change it if you really want this. ------------------------------------------------------------------------- Scott Wood Zoom2 base > +#define NAND_NO_RB 1 > > +#define CONFIG_SYS_NAND_WP More legacy NAND stuff. You should probably go through the config and clean out anything that isn't verifiably necessary. Tom: I Cleaned out a lot more of the config file.

1 0

Re: [U-Boot] Help!Some memory doesn't work on PPC405Ex based board!
by Stefan Roese 14 Apr '09

14 Apr '09

Hi Sun, On Tuesday 14 April 2009, SunNeo wrote: > My platform uses the MICRON MT47H256M8THN DDRII SDRAM and the DDRII SDRAM > is soldered on the board. > > As I said, my board was similar with "Kilauea" evb, so I created my > configuration header file from Kilauea's at U-Boot. In the configuration > file, register value for the DDR SDRAM controller is defined. But I have > removed DDR autocalibraton related configuration from the configuration > file, do you think this will cause any issues? Yes, this could definitely cause some SDRAM related problems. But this is off-topic here. We should continue this discussion on the U-Boot mailing list instead (on CC). So please when you reply to this mail, remove the linuxppc-dev list from the recipients list. Best regards, Stefan

1 0

[U-Boot] [PATCH 1/1] at91sam9/at91cap: fix CONFIG_SYS_HZ to 1000
by Jean-Christophe PLAGNIOL-VILLARD 14 Apr '09

14 Apr '09

The timer has been rewrote with a precision at ~0,18% Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj(a)jcrosoft.com> --- cpu/arm926ejs/at91/timer.c | 82 +++++++++++++++++++++------------------ include/configs/afeb9260.h | 2 +- include/configs/at91cap9adk.h | 2 +- include/configs/at91sam9260ek.h | 2 +- include/configs/at91sam9261ek.h | 2 +- include/configs/at91sam9263ek.h | 2 +- include/configs/at91sam9rlek.h | 2 +- include/configs/pm9263.h | 2 +- 8 files changed, 51 insertions(+), 45 deletions(-) diff --git a/cpu/arm926ejs/at91/timer.c b/cpu/arm926ejs/at91/timer.c index fec545b..c84cb5e 100644 --- a/cpu/arm926ejs/at91/timer.c +++ b/cpu/arm926ejs/at91/timer.c @@ -27,7 +27,9 @@ #include <asm/arch/at91_pit.h> #include <asm/arch/at91_pmc.h> #include <asm/arch/at91_rstc.h> +#include <asm/arch/clk.h> #include <asm/arch/io.h> +#include <div64.h> /* * We're using the AT91CAP9/SAM9 PITC in 32 bit mode, by @@ -36,11 +38,26 @@ #define TIMER_LOAD_VAL 0xfffff #define READ_RESET_TIMER at91_sys_read(AT91_PIT_PIVR) #define READ_TIMER at91_sys_read(AT91_PIT_PIIR) -#define TIMER_FREQ (AT91C_MASTER_CLOCK << 4) -#define TICKS_TO_USEC(ticks) ((ticks) / 6) -ulong get_timer_masked(void); -ulong resettime; +static ulong timestamp; +static ulong lastinc; +static ulong timer_freq; + +static inline unsigned long long tick_to_time(unsigned long long tick) +{ + tick *= CONFIG_SYS_HZ; + do_div(tick, timer_freq); + + return tick; +} + +static inline unsigned long long usec_to_tick(unsigned long long usec) +{ + usec *= timer_freq; + do_div(usec, 1000000); + + return usec; +} /* nothing really to do with interrupts, just starts up a counter. */ int timer_init(void) @@ -56,41 +73,49 @@ int timer_init(void) reset_timer_masked(); + timer_freq = get_mck_clk_rate() >> 4; + return 0; } /* * timer without interrupts */ - -static inline ulong get_timer_raw(void) +unsigned long long get_ticks(void) { ulong now = READ_TIMER; - if (now >= resettime) - return now - resettime; - else - return 0xFFFFFFFFUL - (resettime - now) ; + if (now >= lastinc) /* normal mode (non roll) */ + /* move stamp forward with absolut diff ticks */ + timestamp += (now - lastinc); + else /* we have rollover of incrementer */ + timestamp += (0xFFFFFFFF - lastinc) + now; + lastinc = now; + return timestamp; } void reset_timer_masked(void) { - resettime = READ_TIMER; + /* reset time */ + lastinc = READ_TIMER; /* capture current incrementer value time */ + timestamp = 0; /* start "advancing" time stamp from 0 */ } ulong get_timer_masked(void) { - return TICKS_TO_USEC(get_timer_raw()); - + return tick_to_time(get_ticks()); } -void udelay_masked(unsigned long usec) +void udelay(unsigned long usec) { - ulong tmp; + unsigned long long tmp; + ulong tmo; - tmp = get_timer(0); - while (get_timer(tmp) < usec) /* our timer works in usecs */ - ; /* NOP */ + tmo = usec_to_tick(usec); + tmp = get_ticks() + tmo; /* get current timestamp */ + + while (get_ticks() < tmp) /* loop till event */ + /*NOP*/; } void reset_timer(void) @@ -100,26 +125,7 @@ void reset_timer(void) ulong get_timer(ulong base) { - ulong now = get_timer_masked(); - - if (now >= base) - return now - base; - else - return TICKS_TO_USEC(0xFFFFFFFFUL) - (base - now) ; -} - -void udelay(unsigned long usec) -{ - udelay_masked(usec); -} - -/* - * This function is derived from PowerPC code (read timebase as long long). - * On ARM it just returns the timer value. - */ -unsigned long long get_ticks(void) -{ - return get_timer(0); + return get_timer_masked () - base; } /* diff --git a/include/configs/afeb9260.h b/include/configs/afeb9260.h index ffa7fad..5679b3f 100644 --- a/include/configs/afeb9260.h +++ b/include/configs/afeb9260.h @@ -28,7 +28,7 @@ /* ARM asynchronous clock */ #define AT91_MAIN_CLOCK 18429952 /* from 18.432 MHz crystal */ -#define CONFIG_SYS_HZ 1000000 /* 1us resolution */ +#define CONFIG_SYS_HZ 1000 #define CONFIG_AT91SAM9260 1 /* It's an Atmel AT91SAM9260 SoC*/ #define CONFIG_AFEB9260 1 /* on an AFEB9260 Board */ diff --git a/include/configs/at91cap9adk.h b/include/configs/at91cap9adk.h index 51f9051..b14cd73 100644 --- a/include/configs/at91cap9adk.h +++ b/include/configs/at91cap9adk.h @@ -30,7 +30,7 @@ /* ARM asynchronous clock */ #define AT91_CPU_NAME "AT91CAP9" #define AT91_MAIN_CLOCK 12000000 /* 12 MHz crystal */ -#define CONFIG_SYS_HZ 1000000 /* 1us resolution */ +#define CONFIG_SYS_HZ 1000 #define CONFIG_ARM926EJS 1 /* This is an ARM926EJS Core */ #define CONFIG_AT91CAP9 1 /* It's an Atmel AT91CAP9 SoC */ diff --git a/include/configs/at91sam9260ek.h b/include/configs/at91sam9260ek.h index 558f648..70ceabf 100644 --- a/include/configs/at91sam9260ek.h +++ b/include/configs/at91sam9260ek.h @@ -29,7 +29,7 @@ /* ARM asynchronous clock */ #define AT91_MAIN_CLOCK 18432000 /* 18.432 MHz crystal */ -#define CONFIG_SYS_HZ 1000000 /* 1us resolution */ +#define CONFIG_SYS_HZ 1000 #define CONFIG_ARM926EJS 1 /* This is an ARM926EJS Core */ diff --git a/include/configs/at91sam9261ek.h b/include/configs/at91sam9261ek.h index 5f7bef8..7766513 100644 --- a/include/configs/at91sam9261ek.h +++ b/include/configs/at91sam9261ek.h @@ -30,7 +30,7 @@ /* ARM asynchronous clock */ #define AT91_CPU_NAME "AT91SAM9261" #define AT91_MAIN_CLOCK 18432000 /* 18.432 MHz crystal */ -#define CONFIG_SYS_HZ 1000000 /* 1us resolution */ +#define CONFIG_SYS_HZ 1000 #define CONFIG_ARM926EJS 1 /* This is an ARM926EJS Core */ #define CONFIG_AT91SAM9261 1 /* It's an Atmel AT91SAM9261 SoC*/ diff --git a/include/configs/at91sam9263ek.h b/include/configs/at91sam9263ek.h index 506295b..9119e53 100644 --- a/include/configs/at91sam9263ek.h +++ b/include/configs/at91sam9263ek.h @@ -30,7 +30,7 @@ /* ARM asynchronous clock */ #define AT91_CPU_NAME "AT91SAM9263" #define AT91_MAIN_CLOCK 16367660 /* 16.367 MHz crystal */ -#define CONFIG_SYS_HZ 1000000 /* 1us resolution */ +#define CONFIG_SYS_HZ 1000 #define CONFIG_ARM926EJS 1 /* This is an ARM926EJS Core */ #define CONFIG_AT91SAM9263 1 /* It's an Atmel AT91SAM9263 SoC*/ diff --git a/include/configs/at91sam9rlek.h b/include/configs/at91sam9rlek.h index 8a35303..1aaf9fb 100644 --- a/include/configs/at91sam9rlek.h +++ b/include/configs/at91sam9rlek.h @@ -30,7 +30,7 @@ /* ARM asynchronous clock */ #define AT91_CPU_NAME "AT91SAM9RL" #define AT91_MAIN_CLOCK 12000000 /* 12 MHz crystal */ -#define CONFIG_SYS_HZ 1000000 /* 1us resolution */ +#define CONFIG_SYS_HZ 1000 #define CONFIG_ARM926EJS 1 /* This is an ARM926EJS Core */ #define CONFIG_AT91SAM9RL 1 /* It's an Atmel AT91SAM9RL SoC*/ diff --git a/include/configs/pm9263.h b/include/configs/pm9263.h index 33e320c..5ab8f3f 100644 --- a/include/configs/pm9263.h +++ b/include/configs/pm9263.h @@ -38,7 +38,7 @@ #define MAIN_PLL_DIV 2 /* 2 or 4 */ #define AT91_MAIN_CLOCK 18432000 -#define CONFIG_SYS_HZ 1000000 +#define CONFIG_SYS_HZ 1000 #define CONFIG_ARM926EJS 1 /* This is an ARM926EJS Core */ #define CONFIG_AT91SAM9263 1 /* It's an Atmel AT91SAM9263 SoC*/ -- 1.6.1.3

3 5

[U-Boot] how to build u-boot environment
by ong 14 Apr '09

14 Apr '09

Hi, May i know how to build the u-boot environment for my board? Whats the different below. I would like to know how to build u-boot environment. As i know its different with u-boot.bin. make distclean make at91sam9263ek_config make CROSS_COMPILE=<path_to_cross-compiler/cross-compiler-prefix-> make distclean make at91sam9263ek_dataflash_cs0_config make CROSS_COMPILE=<path_to_cross-compiler/cross-compiler-prefix-> /Ong

2 1

[U-Boot] udelay/udelay_masked problems
by E Robertson 14 Apr '09

14 Apr '09

Hi all, Is their a problem using udelay in board_init functions? When I do this on my at91, arm926ejs board, it seems to be in an endless loop. I thought this was a timer issue but I only have this problem with board init and not in the drivers. Does anyone have any ideas? Is timer_init a arch specific call or should it be board specific for arm?

3 2

[U-Boot] PATCH mtd CFI flash: timeout calculation underflow if imprecise 1kHz timer: fix
by Renato Andreola 14 Apr '09

14 Apr '09

From 6cc63851e5e4c16012d9afad70cedb41b7340de7 Mon Sep 17 00:00:00 2001 From: Renato Andreola <renato.andreola(a)imagos.it> Date: Fri, 10 Apr 2009 12:52:55 +0200 Subject: drivers/mtd/cfi_flash: precision and underflow problem in tout calculation With old configuration it could happen tout=0 if CONFIG_SYS_HZ<1000 solved using an unsigned long long --- drivers/mtd/cfi_flash.c | 4 +++- 1 files changed, 3 insertions(+), 1 deletions(-) diff --git a/drivers/mtd/cfi_flash.c b/drivers/mtd/cfi_flash.c index 175d82a..0aa42a2 100644 --- a/drivers/mtd/cfi_flash.c +++ b/drivers/mtd/cfi_flash.c @@ -660,7 +660,9 @@ static int flash_status_check (flash_info_t * info, flash_sect_t sector, ulong start; #if CONFIG_SYS_HZ != 1000 - tout *= CONFIG_SYS_HZ/1000; + unsigned long long ull; + ull = tout*CONFIG_SYS_HZ + CONFIG_SYS_HZ/2; + tout = ull/1000; /* Compute: tout *= CONFIG_SYS_HZ/1000; */ #endif /* Wait for command completion */ -- 1.5.5

2 1

[U-Boot] [PATCH 1/2] at91: Improved Atmel AT45DB081 dataflash support
by Daniel Gorsulowski 14 Apr '09

14 Apr '09

This patch improves Atmel AT45DB081 dataflash support. Signed-off-by: Daniel Gorsulowski <Daniel.Gorsulowski(a)esd.eu> --- drivers/mtd/dataflash.c | 7 ++++--- 1 files changed, 4 insertions(+), 3 deletions(-) diff --git a/drivers/mtd/dataflash.c b/drivers/mtd/dataflash.c index 96cd395..4b768e7 100644 --- a/drivers/mtd/dataflash.c +++ b/drivers/mtd/dataflash.c @@ -199,7 +199,7 @@ void dataflash_print_info (void) for (i = 0; i < CONFIG_SYS_MAX_DATAFLASH_BANKS; i++) { if (dataflash_info[i].id != 0) { - printf("DataFlash:"); + printf("DataFlash: "); switch (dataflash_info[i].id) { case AT45DB021: printf("AT45DB021\n"); @@ -207,11 +207,12 @@ void dataflash_print_info (void) case AT45DB161: printf("AT45DB161\n"); break; - + case AT45DB081: + printf("AT45DB081\n"); + break; case AT45DB321: printf("AT45DB321\n"); break; - case AT45DB642: printf("AT45DB642\n"); break; -- 1.6.1

3 8

[U-Boot] Expletives in the source code.
by Steve Reckamp 13 Apr '09

13 Apr '09

In looking for offensive language in our source code, we came across several occurrences in u-boot code: \u-boot-1.3.3\board\netphone\phone_console.c Line 987 \u-boot-1.3.3\tools\gdb\ serial.c Line 113 \u-boot-1.3.3\lib_generic\vsprintf.c Line 9 \u-boot-1.3.3\board\w7o\flash.c Line 274 \u-boot-1.3.3\drivers\net\sk98lin\h\skgepnm2.h Line 125 Thanks, Steve Reckamp

7 6