[U-Boot] [PATCH 1/1] m28evk board config

[lothar at denx.de]

Hello Mailing List - Hello Fabio and Marex
Thank you very much for your helpfull hints!!
Just some words of mine on your comments.


Am 2013-09-09 16:37, schrieb Marek Vasut:
> Dear lothar at denx.de,
> 
>>  From 24b6381162b4569ab86b481b8714d81877231f22 Mon Sep 17 00:00:00 
>> 2001
>>  From: Lothar Rubusch <lothar at denx.de>
>> Date: Fri, 6 Sep 2013 15:01:39 +0200
>> Subject: [PATCH] m28evk board specific configurations for setup with 
>> ext
>> boot
>>   partition and separate ext rootfs
>> 
>> 
>> Signed-off-by: Lothar Rubusch <lothar at denx.de>
>> ---
>>   include/configs/m28evk.h | 78
>> ++++++++++++++++++++++++++++++++++++++++++++++--
>>   1 file changed, 75 insertions(+), 3 deletions(-)
>> 
>> diff --git a/include/configs/m28evk.h b/include/configs/m28evk.h
>> index eba8759..b65456e 100644
>> --- a/include/configs/m28evk.h
>> +++ b/include/configs/m28evk.h
>> @@ -150,15 +150,25 @@
>>   #endif
>> 
>>   /* Booting Linux */
>> -#define CONFIG_BOOTDELAY	3
>> +#define CONFIG_BOOTDELAY	5
>>   #define CONFIG_BOOTFILE		"uImage"
>>   #define CONFIG_BOOTARGS		"console=ttyAMA0,115200n8 "
>> -#define CONFIG_BOOTCOMMAND	"run bootcmd_net"
>> +#define CONFIG_BOOTCOMMAND	"run mmc_mmc"
>>   #define CONFIG_LOADADDR		0x42000000
>>   #define CONFIG_SYS_LOAD_ADDR	CONFIG_LOADADDR
> 
> These two options above do not seem right, they certainly can make use 
> of some
> unification. But that's for another patch.

As Fabio Estevan already mentioned, the BOOTDELAY is definitely an 
unnecessary change which I'll leave out.
Fabio further complained about a missing commit history. I can see the 
point, but the configurations were elaborated directly through the uboot 
environment, and thus commited altogether. They were necessary, since 
the before include/configs/m28evk.h was definitely not working well with 
the m28evk board. Testing it, and with a huge help of Marek, I came to 
present this solution, and will soon post a v2 of it.

As Marek explained me personally, his comment here was related to the 
duplicate setting of the LOADADDR. My patch does not address this issue.


>>   /* Extra Environment */
>>   #define CONFIG_EXTRA_ENV_SETTINGS					\
>> +	"fdtfile=imx28-m28evk.dtb\0"                                    \
>> +	"consdev=ttyAMA0\0"                                             \
>> +	"baudrate=115200\0"                                             \
>> +	"bootdev=/dev/mmcblk0p2\0"                                      \
>> +	"rootdev=/dev/mmcblk0p3\0"                                      \
>> +	"netdev=eth0\0"                                                 \
>> +	"hostname=m28evk\0"                                             \
>> +	"rootpath=/opt/eldk-5.3/armv5te/rootfs-qte-sdk\0"               \
> 
> Use 5.4 here ?

Definitely!

>> +	"kernel_addr_r=0x42000000\0"                                    \
>> +	"fdt_addr_r=0x41000000\0"                                       \
>>   	"update_nand_full_filename=u-boot.nand\0"			\
>>   	"update_nand_firmware_filename=u-boot.sb\0"			\
>>   	"update_sd_firmware_filename=u-boot.sd\0"			\
>> @@ -196,7 +206,69 @@
>>   		"setexpr fw_sz ${fw_sz} + 1 ; "				\
>>   		"mmc write ${loadaddr} 0x800 ${fw_sz} ; "		\
>>   		"fi ; "							\
>> -		"fi\0"
>> +		"fi\0"                                                  \
>> +	"addcons=setenv bootargs ${bootargs} 
>> console=${consdev},${baudrate}\0"
>> \
>> +	"addip="                                                        \
>> +		"setenv bootargs ${bootargs} "                          \
>> +			"ip=${ipaddr}:${serverip}:${gatewayip}:${netmask}:" \
>> +				"${hostname}:${netdev}:off\0"           \
>> +	"addmisc=setenv bootargs ${bootargs} ${miscargs}\0"             \
>> +	"adddfltmtd="                                                   \
>> +		"if test \"x${mtdparts}\" == \"x\" ; "                  \
>> +		"then mtdparts default ; "                              \
>> +		"fi\0"                                                  \
> 
> I suspect the addmtd should be called unconditionally for every boot 
> type
> (mmc_nfs...net_nand).

I can't currently see her which conditional case under which addmtd is 
running. I'll double check it for a v2. Thank you.

BR,
Lothar Rubusch