[PATCH 3/6] soc: mediatek: pwrap: Add support for MT6735 and MT6328 SoC/PMIC pair

[AngeloGioacchino Del Regno]

Il 18/10/24 10:10, Yassine Oudjana ha scritto:
> From: Yassine Oudjana <y.oudjana at protonmail.com>
> 
> Add register definitions and configuration for the MT6735 SoC and the
> MT6328 PMIC which are commonly paired and communicate through the PMIC
> wrapper.
> 
> Note that the PMIC wrapper on MT6735M has a slightly different register
> map and is therefore NOT compatible with MT6735.
> 
> Signed-off-by: Yassine Oudjana <y.oudjana at protonmail.com>
> ---
>   drivers/soc/mediatek/mtk-pmic-wrap.c | 251 ++++++++++++++++++++++++++-
>   1 file changed, 248 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/soc/mediatek/mtk-pmic-wrap.c b/drivers/soc/mediatek/mtk-pmic-wrap.c
> index 9fdc0ef792026..b9e8dd2a5999d 100644
> --- a/drivers/soc/mediatek/mtk-pmic-wrap.c
> +++ b/drivers/soc/mediatek/mtk-pmic-wrap.c
> @@ -3,6 +3,7 @@
>    * Copyright (c) 2014 MediaTek Inc.
>    * Author: Flora Fu, MediaTek
>    */
> +
>   #include <linux/clk.h>
>   #include <linux/interrupt.h>
>   #include <linux/io.h>
> @@ -100,7 +101,7 @@ enum dew_regs {
>   	PWRAP_DEW_CIPHER_MODE,
>   	PWRAP_DEW_CIPHER_SWRST,
>   
> -	/* MT6323 only regs */
> +	/* MT6323 and MT6328 only regs */
>   	PWRAP_DEW_CIPHER_EN,
>   	PWRAP_DEW_RDDMY_NO,
>   
> @@ -121,8 +122,10 @@ enum dew_regs {
>   	PWRAP_RG_SPI_CON13,
>   	PWRAP_SPISLV_KEY,
>   
> -	/* MT6359 only regs */
> +	/* MT6359 and MT6328 only regs */
>   	PWRAP_DEW_CRC_SWRST,
> +
> +	/* MT6359 only regs */
>   	PWRAP_DEW_RG_EN_RECORD,
>   	PWRAP_DEW_RECORD_CMD0,
>   	PWRAP_DEW_RECORD_CMD1,
> @@ -171,6 +174,23 @@ static const u32 mt6323_regs[] = {
>   	[PWRAP_DEW_RDDMY_NO] =		0x01a4,
>   };
>   
> +static const u32 mt6328_regs[] = {
> +	[PWRAP_DEW_DIO_EN] =		0x02d4,
> +	[PWRAP_DEW_READ_TEST] =		0x02d6,
> +	[PWRAP_DEW_WRITE_TEST] =	0x02d8,
> +	[PWRAP_DEW_CRC_SWRST] =		0x02da,
> +	[PWRAP_DEW_CRC_EN] =		0x02dc,
> +	[PWRAP_DEW_CRC_VAL] =		0x02de,
> +	[PWRAP_DEW_MON_GRP_SEL] =	0x02e0,
> +	[PWRAP_DEW_CIPHER_KEY_SEL] =	0x02e2,
> +	[PWRAP_DEW_CIPHER_IV_SEL] =	0x02e4,
> +	[PWRAP_DEW_CIPHER_EN] =		0x02e6,
> +	[PWRAP_DEW_CIPHER_RDY] =	0x02e8,
> +	[PWRAP_DEW_CIPHER_MODE] =	0x02ea,
> +	[PWRAP_DEW_CIPHER_SWRST] =	0x02ec,
> +	[PWRAP_DEW_RDDMY_NO] =		0x02ee,
> +};
> +
>   static const u32 mt6331_regs[] = {
>   	[PWRAP_DEW_DIO_EN] =		0x018c,
>   	[PWRAP_DEW_READ_TEST] =		0x018e,
> @@ -394,7 +414,7 @@ enum pwrap_regs {
>   	PWRAP_ADC_RDATA_ADDR1,
>   	PWRAP_ADC_RDATA_ADDR2,
>   
> -	/* MT7622 only regs */
> +	/* MT7622 and MT6735 only regs */
>   	PWRAP_STA,
>   	PWRAP_CLR,
>   	PWRAP_DVFS_ADR8,
> @@ -417,6 +437,8 @@ enum pwrap_regs {
>   	PWRAP_ADC_RDATA_ADDR,
>   	PWRAP_GPS_STA,
>   	PWRAP_SW_RST,
> +
> +	/* MT7622 only regs */
>   	PWRAP_DVFS_STEP_CTRL0,
>   	PWRAP_DVFS_STEP_CTRL1,
>   	PWRAP_DVFS_STEP_CTRL2,
> @@ -481,6 +503,50 @@ enum pwrap_regs {
>   	/* MT8516 only regs */
>   	PWRAP_OP_TYPE,
>   	PWRAP_MSB_FIRST,
> +
> +	/* MT6735 only regs */
> +	PWRAP_WACS3_EN,
> +	PWRAP_INIT_DONE3,
> +	PWRAP_WACS3_CMD,
> +	PWRAP_WACS3_RDATA,
> +	PWRAP_WACS3_VLDCLR,

Are you sure that you need the PWRAP_MD_ADC_xxxx registers in here?

Since MD is relatively big on its own, I'm not sure how to proceed here.. it may
make sense to split the MD part to a different array, or it may not... I do need
to understand what's going on.

Can you please point me at some reference code to look at, so that I can understand
the situation a bit?

Besides, I'm noticing that the "MD_ADC_RDATA_ADDR_R(x)" are sequential registers
and that's on more than just MT6735: instead of having 32 more entries, it might
make more sense to set only the first and declare the number (or size) of regs...

Something like:

enum pwrap_regs {
	.....
	PWRAP_MD_ADC_RDATA_ADDR_LATEST,
	PWRAP_MD_ADC_RDATA_ADDR_WP,
	PWRAP_MD_ADC_RDATA_ADDR_R0,
	PWRAP_MD_ADC_STA0,
	PWRAP_MD_ADC_STA1,
	PWRAP_MD_ADC_STA2
};

static const struct pmic_wrapper_type pwrap_mt6735 = {
	.regs = mt6735_regs,
	.num_md_addr = 32,
	[other stuff]
};

...but again, please, if you can point me at an implementation that actually
uses the R(x) registers, that'd be better ... so that we can choose the best
option to add those in there.

Everything else is great: good job :-)

Cheers,
Angelo