Re: [PATCH 09/11] power: supply: qcom_smbx: add smb5 support

[Luca Weiss]

Hi Casey!

Adding a note here, I also plan to look into what (if any) changes are
necessary for this to work on PMI632 (which is the PMIC for
sdm632/msm8953 Fairphone 3) since that's also SMB5.

On Thu Jun 19, 2025 at 4:55 PM CEST, Casey Connolly wrote:
> Introduce support for the SMB5 charger found on pm8150b and other more
> modern Qualcomm SoCs.
>
> SMB5 is largely similar to SMB2, with a few register differences. The
> main difference is the new Type-C hardware block which some registers
> are moved to.
>
> Signed-off-by: Casey Connolly <casey.connolly@xxxxxxxxxx>
> ---
>  drivers/power/supply/qcom_smbx.c | 367 +++++++++++++++++++++++++++++++++------
>  1 file changed, 314 insertions(+), 53 deletions(-)

<snip>

> +/* Bits 2:0 match PMI8998 TYPE_C_INTRPT_ENB_SOFTWARE_CTRL */
> +#define SMB5_TYPE_C_MODE_CFG				0x544
> +#define SMB5_EN_TRY_SNK_BIT				BIT(4)
> +#define SMB5_EN_SNK_ONLY_BIT				BIT(1)
> +
> +#define SMB5_TYPEC_TYPE_C_VCONN_CONTROL			0x546
> +#define SMB5_VCONN_EN_ORIENTATION_BIT			BIT(2)
> +#define SMB5_VCONN_EN_VALUE_BIT				BIT(1)
> +#define SMB5_VCONN_EN_SRC_BIT				BIT(0)
> +
> +
> +#define SMB5_TYPE_C_DEBUG_ACCESS_SINK			0x54a
> +#define SMB5_TYPEC_DEBUG_ACCESS_SINK_MASK		GENMASK(4, 0)
> +
> +#define SMB5_DEBUG_ACCESS_SRC_CFG			0x54C
> +#define SMB5_EN_UNORIENTED_DEBUG_ACCESS_SRC_BIT	BIT(0)
> +
> +#define SMB5_TYPE_C_EXIT_STATE_CFG			0x550
> +#define SMB5_BYPASS_VSAFE0V_DURING_ROLE_SWAP_BIT	BIT(3)
> +#define SMB5_SEL_SRC_UPPER_REF_BIT			BIT(2)
> +#define SMB5_EXIT_SNK_BASED_ON_CC_BIT			BIT(0)

<snip>

>  /* Init sequence derived from vendor downstream driver */
> -static const struct smb_init_register smb_init_seq[] = {
> -	{ .addr = AICL_RERUN_TIME_CFG, .mask = AICL_RERUN_TIME_MASK, .val = 0 },
> +static const struct smb_init_register smb5_init_seq[] = {
> +	{ .addr = USBIN_CMD_IL, .mask = USBIN_SUSPEND_BIT, .val = 0 },
> +	/*
> +	 * By default configure us as an upstream facing port
> +	 * FIXME: This will be handled by the type-c driver
> +	 */
> +	{ .addr = SMB5_TYPE_C_MODE_CFG,
> +	  .mask = SMB5_EN_TRY_SNK_BIT | SMB5_EN_SNK_ONLY_BIT,
> +	  .val = SMB5_EN_TRY_SNK_BIT },

Since there's already a driver for the Type-C in pm8150b and pm7250b,
can we remove this? Or is additional plumbing between the two drivers
necessary to make this work? Maybe Bryan can also jump in here.

Regards
Luca

> +	{ .addr = SMB5_TYPEC_TYPE_C_VCONN_CONTROL,
> +	  .mask = SMB5_VCONN_EN_ORIENTATION_BIT | SMB5_VCONN_EN_SRC_BIT |
> +		  SMB5_VCONN_EN_VALUE_BIT,
> +	  .val = SMB2_VCONN_EN_SRC_BIT },
> +	{ .addr = SMB5_DEBUG_ACCESS_SRC_CFG,
> +	  .mask = SMB5_EN_UNORIENTED_DEBUG_ACCESS_SRC_BIT,
> +	  .val = SMB5_EN_UNORIENTED_DEBUG_ACCESS_SRC_BIT },
> +	{ .addr = SMB5_TYPE_C_EXIT_STATE_CFG,
> +	  .mask = SMB5_SEL_SRC_UPPER_REF_BIT,
> +	  .val = SMB5_SEL_SRC_UPPER_REF_BIT },
> +	/*
> +	 * Disable Type-C factory mode and stay in Attached.SRC state when VCONN
> +	 * over-current happens
> +	 */
> +	{ .addr = TYPE_C_CFG,
> +	  .mask = APSD_START_ON_CC_BIT,
> +	  .val = 0 },
> +	{ .addr = SMB5_TYPE_C_DEBUG_ACCESS_SINK,
> +	  .mask = SMB5_TYPEC_DEBUG_ACCESS_SINK_MASK,
> +	  .val = 0x17 },