Re: [PATCH] arm64: dts: ti: k3-am62a7-sk: Disable mmc Schmitt Trigger

["Markus Schneider-Pargmann" <msp@baylibre.com>]

[-- Attachment #1: Type: text/plain, Size: 5957 bytes --]

Hi Alexander,

On Tue Jan 6, 2026 at 6:25 PM CET, Alexander Sverdlin wrote:
> Hi Markus,
>
> I'm sorry my patch has caused regression for your use-case!
>
> I think we would need to discuss this with TI via our FAE, because the change
> in question has both been discussed with former FAE and the technical team
> behind, and adopted in TI SDK.
>
> Or have you already discused this with corresponding TI HW team?
>
> Which hardware is affected, is it the official SK-AM62A-LP?
> Is MMC2 the SD-card?

I only tested my am62a board on u-boot v2026.01. It is a SK-AM62A-LP.
MMC2 is the SD-card and mmc1 in the devicetree.

I am using u-boot's am62ax_evm_r5_defconfig and am62ax_evm_a53_defconfig
as defconfigs.

>
> On Tue, 2026-01-06 at 17:22 +0100, Markus Schneider-Pargmann (TI.com) wrote:
>> Remove Schmitt Trigger from mmc pins. With Schmitt Trigger enabled
>> u-boot SPL is not able to read u-boot from mmc:
>> 
>>     Trying to boot from MMC2
>>     Error reading cluster
>>     spl_load_image_fat: error reading image u-boot.img, err - -22
>>     Error: -22
>>     SPL: Unsupported Boot Device!
>>     SPL: failed to boot from all boot devices
>>     ### ERROR ### Please RESET the board ###
>> 
>> I bisected this issue between u-boot v2025.10 and v2026.01 and found the
>> devicetree merge to be the problem. At a closer look I found the
>> k3-pinctrl.h changes. Disabling the Schmitt Trigger fixes the u-boot SPL
>> failure to read from mmc.
>> 
>> Fixes: 5b272127884b ("arm64: dts: ti: k3-pinctrl: Enable Schmitt Trigger by default")
>> Signed-off-by: Markus Schneider-Pargmann (TI.com) <msp@baylibre.com>
>> ---
>>  arch/arm64/boot/dts/ti/k3-am62a7-sk.dts | 36 ++++++++++++++++-----------------
>>  1 file changed, 18 insertions(+), 18 deletions(-)
>> 
>> diff --git a/arch/arm64/boot/dts/ti/k3-am62a7-sk.dts b/arch/arm64/boot/dts/ti/k3-am62a7-sk.dts
>> index e99bdbc2e0cbdf858f1631096f9c2a086191bab3..9129045c8bbd3a83dba6ff6f2148a3624b91b546 100644
>> --- a/arch/arm64/boot/dts/ti/k3-am62a7-sk.dts
>> +++ b/arch/arm64/boot/dts/ti/k3-am62a7-sk.dts
>> @@ -315,30 +315,30 @@ AM62AX_IOPAD(0x0b4, PIN_INPUT_PULLUP, 1) /* (K24) GPMC0_CSn3.I2C2_SDA */
>>  
>>  	main_mmc0_pins_default: main-mmc0-default-pins {
>>  		pinctrl-single,pins = <
>> -			AM62AX_IOPAD(0x220, PIN_INPUT, 0) /* (Y3) MMC0_CMD */
>> -			AM62AX_IOPAD(0x218, PIN_INPUT, 0) /* (AB1) MMC0_CLKLB */
>> -			AM62AX_IOPAD(0x21c, PIN_INPUT, 0) /* (AB1) MMC0_CLK */
>
> according to datasheet, MMC0_CLK should have address 0x218 and it's the ball AB7.
> MMC0_CLKLB is not present in the datasheet and AB1 is actually VSS. 0x21C address
> is not documented.
>
> Something is not right here...
>
> OK, grepping TRM for CLKLB, one can conclude that 0x21c is actually MMC0_CLKLB.
>
> Could you please try to modify 0x21c address only? Does it solve the boot problem?
>
>> -			AM62AX_IOPAD(0x214, PIN_INPUT, 0) /* (AA2) MMC0_DAT0 */
>> -			AM62AX_IOPAD(0x210, PIN_INPUT_PULLUP, 0) /* (AA1) MMC0_DAT1 */
>> -			AM62AX_IOPAD(0x20c, PIN_INPUT_PULLUP, 0) /* (AA3) MMC0_DAT2 */
>> -			AM62AX_IOPAD(0x208, PIN_INPUT_PULLUP, 0) /* (Y4) MMC0_DAT3 */
>> -			AM62AX_IOPAD(0x204, PIN_INPUT_PULLUP, 0) /* (AB2) MMC0_DAT4 */
>> -			AM62AX_IOPAD(0x200, PIN_INPUT_PULLUP, 0) /* (AC1) MMC0_DAT5 */
>> -			AM62AX_IOPAD(0x1fc, PIN_INPUT_PULLUP, 0) /* (AD2) MMC0_DAT6 */
>> -			AM62AX_IOPAD(0x1f8, PIN_INPUT_PULLUP, 0) /* (AC2) MMC0_DAT7 */
>
> All the rest actually have ST enabled on PoR according to TRM and I suppose BootROM
> would have had hard times booting from the affected MMC device if it would not be
> the correct setting?
>
>> +			AM62AX_IOPAD(0x220, PIN_INPUT_NOST, 0) /* (Y3) MMC0_CMD */
>> +			AM62AX_IOPAD(0x218, PIN_INPUT_NOST, 0) /* (AB1) MMC0_CLKLB */
>> +			AM62AX_IOPAD(0x21c, PIN_INPUT_NOST, 0) /* (AB1) MMC0_CLK */
>> +			AM62AX_IOPAD(0x214, PIN_INPUT_NOST, 0) /* (AA2) MMC0_DAT0 */
>> +			AM62AX_IOPAD(0x210, PIN_INPUT_PULLUP_NOST, 0) /* (AA1) MMC0_DAT1 */
>> +			AM62AX_IOPAD(0x20c, PIN_INPUT_PULLUP_NOST, 0) /* (AA3) MMC0_DAT2 */
>> +			AM62AX_IOPAD(0x208, PIN_INPUT_PULLUP_NOST, 0) /* (Y4) MMC0_DAT3 */
>> +			AM62AX_IOPAD(0x204, PIN_INPUT_PULLUP_NOST, 0) /* (AB2) MMC0_DAT4 */
>> +			AM62AX_IOPAD(0x200, PIN_INPUT_PULLUP_NOST, 0) /* (AC1) MMC0_DAT5 */
>> +			AM62AX_IOPAD(0x1fc, PIN_INPUT_PULLUP_NOST, 0) /* (AD2) MMC0_DAT6 */
>> +			AM62AX_IOPAD(0x1f8, PIN_INPUT_PULLUP_NOST, 0) /* (AC2) MMC0_DAT7 */
>>  		>;
>>  		bootph-all;
>>  	};
>>  
>>  	main_mmc1_pins_default: main-mmc1-default-pins {
>>  		pinctrl-single,pins = <
>> -			AM62AX_IOPAD(0x23c, PIN_INPUT, 0) /* (A21) MMC1_CMD */
>> -			AM62AX_IOPAD(0x234, PIN_INPUT, 0) /* (B22) MMC1_CLK */
>> -			AM62AX_IOPAD(0x230, PIN_INPUT, 0) /* (A22) MMC1_DAT0 */
>> -			AM62AX_IOPAD(0x22c, PIN_INPUT, 0) /* (B21) MMC1_DAT1 */
>> -			AM62AX_IOPAD(0x228, PIN_INPUT, 0) /* (C21) MMC1_DAT2 */
>> -			AM62AX_IOPAD(0x224, PIN_INPUT, 0) /* (D22) MMC1_DAT3 */
>> -			AM62AX_IOPAD(0x240, PIN_INPUT, 0) /* (D17) MMC1_SDCD */
>
> All of these have ST enabled on PoR, according to TRM.
>
>> +			AM62AX_IOPAD(0x23c, PIN_INPUT_NOST, 0) /* (A21) MMC1_CMD */
>> +			AM62AX_IOPAD(0x234, PIN_INPUT_NOST, 0) /* (B22) MMC1_CLK */
>> +			AM62AX_IOPAD(0x230, PIN_INPUT_NOST, 0) /* (A22) MMC1_DAT0 */
>> +			AM62AX_IOPAD(0x22c, PIN_INPUT_NOST, 0) /* (B21) MMC1_DAT1 */
>> +			AM62AX_IOPAD(0x228, PIN_INPUT_NOST, 0) /* (C21) MMC1_DAT2 */
>> +			AM62AX_IOPAD(0x224, PIN_INPUT_NOST, 0) /* (D22) MMC1_DAT3 */
>> +			AM62AX_IOPAD(0x240, PIN_INPUT_NOST, 0) /* (D17) MMC1_SDCD */

My board is setup to boot from SD card for easier u-boot testing. So
only the mmc1 is relevant for my setup. I just tested which pins need
NOST here for the boot to work, all DAT pins need the NOST variants,
otherwise it does not boot here. Not sure if this is just my board or it
fails on other boards as well.

Best
Markus


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 289 bytes --]