On Thu, Aug 28, 2014 at 9:58 PM, Michal Simek <michal.si...@xilinx.com> wrote:
> On 08/28/2014 01:02 PM, Peter Crosthwaite wrote:
>> Ping!
>>
>> On Wed, Aug 20, 2014 at 10:14 PM, Peter Crosthwaite
>> <crosthwaitepe...@gmail.com> wrote:
>>> To better enable debug of u-boot itself (in particular SPL). To debug
>>> u-boot you want to put your board in JTAG boot mode for quick recompile
>>> and elf downloads via the JTAG debugger. Yet you may still want to boot
>>> from a persistent storage media (SD or QSPI). Allow override of the
>>> boot mode pins to facilitate this.
>>>
>>> Signed-off-by: Peter Crosthwaite <crosthwaite.pe...@gmail.com>
>>> ---
>>> arch/arm/cpu/armv7/zynq/slcr.c | 4 ++++
>>> doc/README.zynq | 6 ++++++
>>> 2 files changed, 10 insertions(+)
>>>
>>> diff --git a/arch/arm/cpu/armv7/zynq/slcr.c b/arch/arm/cpu/armv7/zynq/slcr.c
>>> index 934ccc3..26e02b8 100644
>>> --- a/arch/arm/cpu/armv7/zynq/slcr.c
>>> +++ b/arch/arm/cpu/armv7/zynq/slcr.c
>>> @@ -155,8 +155,12 @@ void zynq_slcr_devcfg_enable(void)
>>>
>>> u32 zynq_slcr_get_boot_mode(void)
>>> {
>>> +#ifdef CONFIG_ZYNQ_BM_FORCE
>>> + return CONFIG_ZYNQ_BM_FORCE;
>>> +#else
>>> /* Get the bootmode register value */
>>> return readl(&slcr_base->boot_mode);
>>> +#endif
>>> }
>>>
>>> u32 zynq_slcr_get_idcode(void)
>>> diff --git a/doc/README.zynq b/doc/README.zynq
>>> index 043c970..70be7ae 100644
>>> --- a/doc/README.zynq
>>> +++ b/doc/README.zynq
>>> @@ -52,6 +52,12 @@ board_late_init() will read the bootmode values using
>>> slcr bootmode register
>>> at runtime and assign the modeboot variable to specific bootmode string
>>> which
>>> is intern used in autoboot.
>>>
>>> +This value can be overridden at compile time with the define
>>> +CONFIG_ZYNQ_BM_FORCE. This is useful for debugging a u-boot elf downloaded
>>> +over JTAG. E.g. The board will be jumpered for ZYNQ_BM_JTAG, but you can
>>> +force u-boot to attempt an SD or QSPI boot should that be what you want to
>>> +debug.
>>> +
>
> I understand the purpose of this patch but I don't think it is useful to add
> and support this option.
> Someone else could come with reading GPIO for this purpose, etc which is just
> the same case.
>
> I am happy to add the patch which shows bootmode register content or any ? :
> logic around
> in debug() print but adding new CONFIG_ option is just too much.
> (Also every new config option should be properly documented).
>
I'm not sure what this adds? Where does the debug print get involved?
> The second option is to setup spl_boot_device in arch zynq as weak function
> and overwrite it.
> Then you can setup whatever you like in your board file.
>
Sounds like the winning compromise.
Regards,
Peter
> Thanks,
> Michal
>
_______________________________________________
U-Boot mailing list
U-Boot@lists.denx.de
http://lists.denx.de/mailman/listinfo/u-boot
