[PATCH v2 1/3] ARM: dts: aspeed: mtjade: Add some gpios
Quan Nguyen
quan at os.amperecomputing.com
Sun Oct 24 21:44:04 PDT 2021
On 19/10/2021 13:46, Joel Stanley wrote:
> On Tue, 19 Oct 2021 at 06:02, Quan Nguyen <quan at os.amperecomputing.com> wrote:
>>
>> Add S0_SCP_AUTH_FAIL, S1_SCP_AUTH_FAIL gpios to indicates firmware
>> authentication fail on each socket.
>
> These use the gpio-keys API to expose the GPIOs. I think OpenBMC is
> moving away from this abstraction, and instead reading the GPIOs with
> the gpio chardev interface.
>
Thanks Joel for the review,
Yes, will switch to use gpio chardev interface and will remove these
gpio-keys defined in next version
>>
>> Add gpio RTC_BAT_SEN_EN to enable RTC battery adc sensor.
>>
>> Add BMC_I2C4_O_EN gpio to go high at boot to enable access to I2C4 bus.
>
> OpenBMC has started a process to document GPIOs that are exposed to
> userspace, initially so a common userspace can be used across
> machines. I like doing it for the additional reason that it provides
> consistency in the naming.
>
> https://github.com/openbmc/docs/blob/master/designs/device-tree-gpio-naming.md
>
> If you could take a look at that document and add your GPIOs where
> possible, and then update the device tree.
>
There is on-going discussion about naming of the RTC voltage sensing
enable gpio pin in [1], So I'd like to address this change later in
separate patch when this is settled. Hence, I will remove this changes
in next version.
[1] https://gerrit.openbmc-project.xyz/c/openbmc/docs/+/48018
About the BMC_I2C4_O_EN, this is needed as gpio-hog to make the i2c bus
accessible at boot time so I will keep this in next version.
In fact, I'm still a bit of confuse if this pin needs a common name as
it is just to enable the bus access and does not serve for any
particular common purpose.
Thanks,
- Quan
>>
>> Signed-off-by: Quan Nguyen <quan at os.amperecomputing.com>
>> Signed-off-by: Thang Nguyen <thang at os.amperecomputing.com>
>> ---
>> v2:
>> - None
>>
>> .../arm/boot/dts/aspeed-bmc-ampere-mtjade.dts | 21 ++++++++++++++++++-
>> 1 file changed, 20 insertions(+), 1 deletion(-)
>>
>> diff --git a/arch/arm/boot/dts/aspeed-bmc-ampere-mtjade.dts b/arch/arm/boot/dts/aspeed-bmc-ampere-mtjade.dts
>> index 57b0c45a2298..3515d55bd312 100644
>> --- a/arch/arm/boot/dts/aspeed-bmc-ampere-mtjade.dts
>> +++ b/arch/arm/boot/dts/aspeed-bmc-ampere-mtjade.dts
>> @@ -86,6 +86,18 @@ S0_cpu_fault {
>> linux,code = <ASPEED_GPIO(J, 1)>;
>> };
>>
>> + S0_scp_auth_fail {
>> + label = "S0_SCP_AUTH_FAIL";
>> + gpios = <&gpio ASPEED_GPIO(J, 2) GPIO_ACTIVE_LOW>;
>> + linux,code = <ASPEED_GPIO(J, 2)>;
>> + };
>> +
>> + S1_scp_auth_fail {
>> + label = "S1_SCP_AUTH_FAIL";
>> + gpios = <&gpio ASPEED_GPIO(Z, 5) GPIO_ACTIVE_LOW>;
>> + linux,code = <ASPEED_GPIO(Z, 5)>;
>> + };
>> +
>> S1_overtemp {
>> label = "S1_OVERTEMP";
>> gpios = <&gpio ASPEED_GPIO(Z, 6) GPIO_ACTIVE_LOW>;
>> @@ -590,7 +602,7 @@ &gpio {
>> /*Q0-Q7*/ "","","","","","UID_BUTTON","","",
>> /*R0-R7*/ "","","BMC_EXT_HIGHTEMP_L","OCP_AUX_PWREN",
>> "OCP_MAIN_PWREN","RESET_BUTTON","","",
>> - /*S0-S7*/ "","","","","","","","",
>> + /*S0-S7*/ "","","","","RTC_BAT_SEN_EN","","","",
>
> I suggest you create a proposal to call this one
> battery-voltage-read-enable. I know that some of the IBM machines
> intend to have this same GPIO.
>
>> /*T0-T7*/ "","","","","","","","",
>> /*U0-U7*/ "","","","","","","","",
>> /*V0-V7*/ "","","","","","","","",
>> @@ -604,4 +616,11 @@ &gpio {
>> "S1_BMC_DDR_ADR","","","","",
>> /*AC0-AC7*/ "SYS_PWR_GD","","","","","BMC_READY","SLAVE_PRESENT_L",
>> "BMC_OCP_PG";
>> +
>> + i2c4_o_en {
>> + gpio-hog;
>> + gpios = <ASPEED_GPIO(Y, 2) GPIO_ACTIVE_HIGH>;
>> + output-high;
>> + line-name = "BMC_I2C4_O_EN";
>> + };
>> };
>> --
>> 2.28.0
>>
More information about the linux-arm-kernel
mailing list