[U-Boot] [U-Boot, 1/3] rockchip: boot mode: parse adc channel from dts

[Andy Yan]

Hi Philipp:


On 2017年11月28日 21:59, Philipp Tomsich wrote:
> +sjg
>
> On Tue, 28 Nov 2017, Andy Yan wrote:
>
>> Most the current rockchip based boards use adc channel
>> 1 detect the download key, but there are also some
>> boards like rv1108 based plaform use adc channel 0.
>> So we parse the adc channel from dts if we can get
>> it, otherwise we use the channel 1 as default.
>>
>> Signed-off-by: Andy Yan <andy.yan at rock-chips.com>
>> Acked-by: Philipp Tomsich <philipp.tomsich at theobroma-systems.com>
>> ---
>>
>> arch/arm/mach-rockchip/boot_mode.c | 15 ++++++++++++++-
>> 1 file changed, 14 insertions(+), 1 deletion(-)
>>
>> diff --git a/arch/arm/mach-rockchip/boot_mode.c 
>> b/arch/arm/mach-rockchip/boot_mode.c
>> index 942849f..49dfd39 100644
>> --- a/arch/arm/mach-rockchip/boot_mode.c
>> +++ b/arch/arm/mach-rockchip/boot_mode.c
>> @@ -8,6 +8,9 @@
>> #include <adc.h>
>> #include <asm/io.h>
>> #include <asm/arch/boot_mode.h>
>> +#include <fdtdec.h>
>> +
>> +DECLARE_GLOBAL_DATA_PTR;
>>
>> void set_back_to_bootrom_dnl_flag(void)
>> {
>> @@ -26,9 +29,19 @@ void set_back_to_bootrom_dnl_flag(void)
>>
>> __weak int rockchip_dnl_key_pressed(void)
>> {
>> +    const void *blob = gd->fdt_blob;
>>     unsigned int val;
>> +    int channel = 1;
>> +    int node;
>> +    u32 chns[2];
>> +
>> +    node = fdt_node_offset_by_compatible(blob, 0, "adc-keys");
>> +    if (node >= 0) {
>> +        if (!fdtdec_get_int_array(blob, node, "io-channels", chns, 2))
>> +            channel = chns[1];
>> +    }
>
> The driver for 'adc-keys' should be a driver in drivers/input that can 
> then be retrieved via DM and queried using keyboard_getc().

     Yes, if there is an exiting adc-keys driver, we will use it here 
with no doubts, but there is not now. I just need  to check the button
down event once, no up or other things needed, so I call the adc api 
directly. I grep all the u-boot project, and found all other boards
handle this kind of button status check by call device specific api 
directly, rather than use the api from input subsystem. So I think
this is a acceptable way.
       Would you please consider taking this patch if there is no other 
problems. And when there is a adc-keys driver in the future, I can move
it to input based api, or I will write a adc-keys driver when my 
workloads is not so heavy as now.
>
>>
>> -    if (adc_channel_single_shot("saradc", 1, &val)) {
>> +    if (adc_channel_single_shot("saradc", channel, &val)) {
>>         pr_err("%s: adc_channel_single_shot fail!\n", __func__);
>>         return false;
>>     }
>>
>
>
>