[U-Boot] [PATCH 1/2] x86: x86-common.h: Add CONFIG_BOOTDELAY

[Stefan Roese]

Hi Simon,

On 28.01.2016 17:16, Simon Glass wrote:
> Hi Stefan,
>
> On 28 January 2016 at 09:13, Stefan Roese <sr at denx.de> wrote:
>>
>> Hi Bin,
>>
>> (added Simon to Cc)
>>
>> On 26.01.2016 07:48, Bin Meng wrote:
>>>
>>> On Tue, Jan 19, 2016 at 2:19 PM, Bin Meng <bmeng.cn at gmail.com> wrote:
>>>>
>>>> On Mon, Jan 18, 2016 at 9:49 PM, Stefan Roese <sr at denx.de> wrote:
>>>>>
>>>>> Without this CONFIG_BOOTDELAY, autobooting does not work at all. As
>>>>> autoboot_command() from common/* will not get called. So lets define
>>>>> CONFIG_BOOTDELAY, so that auto-booting works on x86.
>>>>>
>>>>> Signed-off-by: Stefan Roese <sr at denx.de>
>>>>> Cc: Miao Yan <yanmiaobest at gmail.com>
>>>>> Cc: Bin Meng <bmeng.cn at gmail.com>
>>>>> Cc: Simon Glass <sjg at chromium.org>
>>>>> ---
>>>>>    include/configs/x86-common.h | 2 ++
>>>>>    1 file changed, 2 insertions(+)
>>>>>
>>>>
>>>> Acked-by: Bin Meng <bmeng.cn at gmail.com>
>>>
>>>
>>> Sorry, this patch does not build for efi-x86.
>>>
>>>          x86:  +   efi-x86
>>> +../common/autoboot.c: In function 'process_fdt_options':
>>> +../common/autoboot.c:296:36: error: 'CONFIG_SYS_TEXT_BASE' undeclared
>>> (first use in this function)
>>> +   setenv_addr("kernaddr", (void *)(CONFIG_SYS_TEXT_BASE + addr));
>>> +                                    ^
>>> +../common/autoboot.c:296:36: note: each undeclared identifier is
>>> reported only once for each function it appears in
>>> +make[2]: *** [common/autoboot.o] Error 1
>>> +make[1]: *** [common] Error 2
>>> +make: *** [sub-make] Error 2
>>>
>>> Could you please fix this? Sorry I did not run buildman earlier.
>>
>>
>> I'm a bit hesitant on how to fix this. As I don't really know this
>> "efi-x86" target in detail. Is this code in process_fdt_options()
>> really needed for this target? To configure the env variables
>> "kernaddr" and "rootaddr" dynamically from the DT properties
>> "kernel-offset" and "rootdisk-offset". I can't find any references
>> to these DT properties anywhere?
>>
>> Simon, you introduced this env variable handling with the patch
>> [fdt: Set kernaddr if fdt indicates a kernel is present] (git ID
>> fcabc24f) in October 2012.
>>
>> Perhaps its best to assign TEXT_BASE to 0 if its not defined at
>> all? Or is this in general the correct value for the "efi-x86"
>> target and should be set specifically for it?
>
> It is a funny target since we actually produce a relocatable ELF. I'd
> suggest putting #ifdef CONFIG_SYS_TEXT_BASE around it.

Okay. I'll "solve" it this way.

Thanks,
Stefan