[U-Boot] [PATCH v2 01/11] ARM: fix relocation on ARM926EJS
José Miguel Gonçalves
jose.goncalves at inov.pt
Mon Sep 17 11:20:07 CEST 2012
Hi Christian,
On 09/17/2012 10:03 AM, Christian Riesch wrote:
> Hi,
>
> On Mon, Sep 17, 2012 at 10:34 AM, José Miguel Gonçalves
> <jose.goncalves at inov.pt> wrote:
>> On 09/17/2012 07:28 AM, Christian Riesch wrote:
>>> Hi,
>>>
>>> On Sun, Sep 16, 2012 at 5:36 PM, Marek Vasut <marex at denx.de> wrote:
>>>> Dear José Miguel Gonçalves,
>>>>
>>>>> On 09/16/2012 11:06 AM, Marek Vasut wrote:
>>>>>> Dear José Miguel Gonçalves,
>>>>>>
>>>>>>> On 09/15/2012 07:03 PM, Marek Vasut wrote:
>>>>>>>> Dear José Miguel Gonçalves,
>>>>>>>>
>>>>>>>>> Jumping to board_init_r is not performed due to a bug on address
>>>>>>>>> computation.
>>>>>>>> Is your CONFIG_SYS_TEXT_BASE configured correctly? I don't detect any
>>>>>>>> misbehavior on my arm926 boards.
>>>>>>> Maybe because you are not using it to build an SPL?
>>>>>> I do ... and I use CONFIG_SPL_TEXT_BASE properly .
>>>>>>> Please check the same chunk of code in other start.S for arm1176 and
>>>>>>> armv7. They have the same code that I put for arm926ejs.
>>>>>> Please wait and please first explain what is the issue.
>>>>> The issue is what I've explained in the patch comments.
>>>> "Jumping to board_init_r is not performed due to a bug on address
>>>> computation."
>>>>
>>>> Ok, I don't know how to replicate the bug from this comment or what
>>>> effects it
>>>> causes or ... well, anything. So please, try to be more elaborate in your
>>>> patch
>>>> description next time. Anyway ..
>>> Same for me - I have no idea what you are trying to fix here. In my
>>> SPL configuration, _TEXT_BASE and _start point to the same location,
>>> so please explain why they are different on your board.
>>
>> They are different because of how start.S is implemented in arm926ejs.
>>
>> In my SPL map file I see:
>>
>> .text 0x0000000000000000 0xc24
>> arch/arm/cpu/arm926ejs/start.o(.text)
>> .text 0x0000000000000000 0x120 arch/arm/cpu/arm926ejs/start.o
>> 0x0000000000000000 _start
>> 0x0000000000000040 _TEXT_BASE
>> 0x0000000000000044 _bss_start_ofs
>> 0x0000000000000048 _bss_end_ofs
>> 0x000000000000004c _end_ofs
>> 0x0000000000000050 IRQ_STACK_START_IN
>> 0x0000000000000074 relocate_code
> So _start is 0x00000000, and in your
> config/include/include/configs/mini2416.h you have
>
> #define CONFIG_SPL_TEXT_BASE 0x00000000
>
> and in arch/arm/cpu/arm926ejs/start.S we have
>
> .globl _TEXT_BASE
> _TEXT_BASE:
> #ifdef CONFIG_NAND_SPL /* deprecated, use instead CONFIG_SPL_BUILD */
> .word CONFIG_SYS_TEXT_BASE
> #else
> #ifdef CONFIG_SPL_BUILD
> .word CONFIG_SPL_TEXT_BASE
> #else
> .word CONFIG_SYS_TEXT_BASE
> #endif
> #endif
>
> So
>
> ldr r1, _TEXT_BASE
>
> should be the same as
>
> adr r1, _start
>
> However, if you do not load your SPL to 0x00000000 but to another
> address and execute it from there, it will be different, since adr
> uses relative adressing, right? Are you sure you are loading it to
> 0x00000000?
Not an expert on ARM assembly, so cannot give any feedback on the
differences between adr and ldr mnemonics.
What I know for a fact is that the S3C2416, when booting from it's
internal ROM, makes a copy of the first 8KB of the NAND flash to an
internal RAM (named SteppingStone), maps it at address 0 and the jumps
to the start of it. This mapping is not clear in Samsung's user manual
but I retrieved that information from here:
http://barebox.org/documentation/barebox-2011.05.0/dev_s3c24xx_arch.html
>
> Regards, Christian
>
>>
>>
>>>>> Without this
>>>>> change the code never reaches board_init_r in the SPL and I think I have
>>>>> all the configurations correctly set.
>>>> I wonder why you'd ever want to reach board_init_r in the SPL. SPL is
>>>> there only
>>>> to load the real U-Boot from whatever media, so you usually use either
>>>> NAND SPL
>>>> or something like that.
>>>>
>>> Marek, going into board_init_r is fine for SPL, for example the
>>> davinci SPL does some hw initialization in board_init_f and then loads
>>> u-boot in board_init_r.
>>> Regards, Christian
>>>
>>>> What do you boot the rest from ?
>>>>
>>>>> If the bug is not from here please
>>>>> suggest me what I need to change in the configuration in order to
>>>>> correctly boot my board.
>>>>>
>>>>>>>>> Relocation offsets are not needed when building SPL.
>>>>>>>> Do they cause any trouble?
>>>>>>> No! Just not needed.
>>>>>>>
>>>>>>>>> Signed-off-by: José Miguel Gonçalves <jose.goncalves at inov.pt>
>>>>>>>>> ---
>>>>>>>>>
>>>>>>>>> Changes for v2:
>>>>>>>>> - None
>>>>>>>>>
>>>>>>>>> ---
>>>>>>>>>
>>>>>>>>> arch/arm/cpu/arm926ejs/start.S | 4 +++-
>>>>>>>>> 1 file changed, 3 insertions(+), 1 deletion(-)
>>>>>>>>>
>>>>>>>>> diff --git a/arch/arm/cpu/arm926ejs/start.S
>>>>>>>>> b/arch/arm/cpu/arm926ejs/start.S index 6f05f1a..2da5342 100644
>>>>>>>>> --- a/arch/arm/cpu/arm926ejs/start.S
>>>>>>>>> +++ b/arch/arm/cpu/arm926ejs/start.S
>>>>>>>>>
>>>>>>>>> @@ -325,7 +325,7 @@ _nand_boot_ofs:
>>>>>>>>> .word nand_boot
>>>>>>>>>
>>>>>>>>> #else
>>>>>>>>>
>>>>>>>>> ldr r0, _board_init_r_ofs
>>>>>>>>>
>>>>>>>>> - ldr r1, _TEXT_BASE
>>>>>>>>> + adr r1, _start
>>>>>>>>>
>>>>>>>>> add lr, r0, r1
>>>>>>>>> add lr, lr, r9
>>>>>>>>> /* setup parameters for board_init_r */
>>>>>>>>>
>>>>>>>>> @@ -338,12 +338,14 @@ _board_init_r_ofs:
>>>>>>>>> .word board_init_r - _start
>>>>>>>>>
>>>>>>>>> #endif
>>>>>>>>>
>>>>>>>>> +#ifndef CONFIG_SPL_BUILD
>>>>>>>>>
>>>>>>>>> _rel_dyn_start_ofs:
>>>>>>>>> .word __rel_dyn_start - _start
>>>>>>>>>
>>>>>>>>> _rel_dyn_end_ofs:
>>>>>>>>> .word __rel_dyn_end - _start
>>>>>>>>>
>>>>>>>>> _dynsym_start_ofs:
>>>>>>>>> .word __dynsym_start - _start
>>>>>>>>>
>>>>>>>>> +#endif
>>>>>>>>>
>>>>>>>>> /*
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> ******************************************************************
>>>>>>>>> *** ****
>>>>>>>> Best regards,
>>>>>>>> Marek Vasut
>>>>>>> Best regards,
>>>>>>> José Gonçalves
>>>>>> Best regards,
>>>>>> Marek Vasut
>>>>> Best regards,
>>>>> José Gonçalves
>>
>> Best regards,
>> José Gonçalves
>
Best regards,
José Gonçalves
More information about the U-Boot
mailing list