Re: [PATCH v11 4/4] pinctrl: mediatek: add rsel setting on MT8195

From: Chen-Yu Tsai
Date: Mon Sep 06 2021 - 06:09:43 EST

Next message: Ryan Chen: "RE: [PATCH v3] media: aspeed-video: ignore interrupts that aren't enabled"
Previous message: Paolo Bonzini: "Re: [PATCH 1/2] KVM: VMX: avoid running vmx_handle_exit_irqoff in case of emulation"
In reply to: Chen-Yu Tsai: "Re: [PATCH v11 4/4] pinctrl: mediatek: add rsel setting on MT8195"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

On Mon, Sep 6, 2021 at 11:17 AM zhiyong.tao <zhiyong.tao@xxxxxxxxxxxx> wrote:
>
> On Wed, 2021-09-01 at 18:10 +0800, Chen-Yu Tsai wrote:
> > Hi,
> >
> > On Mon, Aug 30, 2021 at 8:37 AM Zhiyong Tao <zhiyong.tao@xxxxxxxxxxxx
> > > wrote:
> > >
> > > This patch provides rsel setting on MT8195
> >
> > A bit more context, like what is rsel, would be nice.
> >
> we will add more context it in the next version.
> > >
> > > Signed-off-by: Zhiyong Tao <zhiyong.tao@xxxxxxxxxxxx>
> > > ---
> > > drivers/pinctrl/mediatek/pinctrl-mt8195.c | 133 +++++++++++++
> > > .../pinctrl/mediatek/pinctrl-mtk-common-v2.c | 181
> > > ++++++++++++++++--
> > > .../pinctrl/mediatek/pinctrl-mtk-common-v2.h | 35 +++-
> > > drivers/pinctrl/mediatek/pinctrl-paris.c | 24 ++-
> > > drivers/pinctrl/mediatek/pinctrl-paris.h | 2 +-
> > > 5 files changed, 348 insertions(+), 27 deletions(-)
> >
> > Could you split this patch into two, so the common parts are in the
> > first
> > patch, and the mt8195 specific changes are in the second?
> >
>
> ok, we will split in the next version.
>
> > >
> > > diff --git a/drivers/pinctrl/mediatek/pinctrl-mt8195.c
> > > b/drivers/pinctrl/mediatek/pinctrl-mt8195.c
> > > index 892e79703f98..275b7ba4386d 100644
> > > --- a/drivers/pinctrl/mediatek/pinctrl-mt8195.c
> > > +++ b/drivers/pinctrl/mediatek/pinctrl-mt8195.c
> > > @@ -779,6 +779,135 @@ static const struct mtk_pin_field_calc
> > > mt8195_pin_drv_adv_range[] = {
> > > PIN_FIELD_BASE(45, 45, 1, 0x040, 0x10, 9, 3),
> > > };
> > >
> > > +static const struct mtk_pin_field_calc mt8195_pin_rsel_range[] = {
> > > + PIN_FIELD_BASE(8, 8, 4, 0x0c0, 0x10, 15, 3),
> > > + PIN_FIELD_BASE(9, 9, 4, 0x0c0, 0x10, 0, 3),
> > > + PIN_FIELD_BASE(10, 10, 4, 0x0c0, 0x10, 18, 3),
> > > + PIN_FIELD_BASE(11, 11, 4, 0x0c0, 0x10, 3, 3),
> > > + PIN_FIELD_BASE(12, 12, 4, 0x0c0, 0x10, 21, 3),
> > > + PIN_FIELD_BASE(13, 13, 4, 0x0c0, 0x10, 6, 3),
> > > + PIN_FIELD_BASE(14, 14, 4, 0x0c0, 0x10, 24, 3),
> > > + PIN_FIELD_BASE(15, 15, 4, 0x0c0, 0x10, 9, 3),
> > > + PIN_FIELD_BASE(16, 16, 4, 0x0c0, 0x10, 27, 3),
> > > + PIN_FIELD_BASE(17, 17, 4, 0x0c0, 0x10, 12, 3),
> > > + PIN_FIELD_BASE(29, 29, 2, 0x080, 0x10, 0, 3),
> > > + PIN_FIELD_BASE(30, 30, 2, 0x080, 0x10, 3, 3),
> > > + PIN_FIELD_BASE(34, 34, 1, 0x0e0, 0x10, 0, 3),
> > > + PIN_FIELD_BASE(35, 35, 1, 0x0e0, 0x10, 3, 3),
> > > + PIN_FIELD_BASE(44, 44, 1, 0x0e0, 0x10, 6, 3),
> > > + PIN_FIELD_BASE(45, 45, 1, 0x0e0, 0x10, 9, 3),
> > > +};
> > > +
> > > +static const struct mtk_pin_rsel mt8195_pin_rsel_val_range[] = {
> > > + PIN_RSEL(8, 17, 0x0, 75000, 75000),
> > > + PIN_RSEL(8, 17, 0x1, 10000, 5000),
> > > + PIN_RSEL(8, 17, 0x2, 5000, 75000),
> > > + PIN_RSEL(8, 17, 0x3, 4000, 5000),
> > > + PIN_RSEL(8, 17, 0x4, 3000, 75000),
> > > + PIN_RSEL(8, 17, 0x5, 2000, 5000),
> > > + PIN_RSEL(8, 17, 0x6, 1500, 75000),
> > > + PIN_RSEL(8, 17, 0x7, 1000, 5000),
> > > + PIN_RSEL(29, 30, 0x0, 75000, 75000),
> > > + PIN_RSEL(29, 30, 0x1, 10000, 5000),
> > > + PIN_RSEL(29, 30, 0x2, 5000, 75000),
> > > + PIN_RSEL(29, 30, 0x3, 4000, 5000),
> > > + PIN_RSEL(29, 30, 0x4, 3000, 75000),
> > > + PIN_RSEL(29, 30, 0x5, 2000, 5000),
> > > + PIN_RSEL(29, 30, 0x6, 1500, 75000),
> > > + PIN_RSEL(29, 30, 0x7, 1000, 5000),
> > > + PIN_RSEL(34, 35, 0x0, 75000, 75000),
> > > + PIN_RSEL(34, 35, 0x1, 10000, 5000),
> > > + PIN_RSEL(34, 35, 0x2, 5000, 75000),
> > > + PIN_RSEL(34, 35, 0x3, 4000, 5000),
> > > + PIN_RSEL(34, 35, 0x4, 3000, 75000),
> > > + PIN_RSEL(34, 35, 0x5, 2000, 5000),
> > > + PIN_RSEL(34, 35, 0x6, 1500, 75000),
> > > + PIN_RSEL(34, 35, 0x7, 1000, 5000),
> > > + PIN_RSEL(44, 45, 0x0, 75000, 75000),
> > > + PIN_RSEL(44, 45, 0x1, 10000, 5000),
> > > + PIN_RSEL(44, 45, 0x2, 5000, 75000),
> > > + PIN_RSEL(44, 45, 0x3, 4000, 5000),
> > > + PIN_RSEL(44, 45, 0x4, 3000, 75000),
> > > + PIN_RSEL(44, 45, 0x5, 2000, 5000),
> > > + PIN_RSEL(44, 45, 0x6, 1500, 75000),
> > > + PIN_RSEL(44, 45, 0x7, 1000, 5000),
> > > +};
> > > +
> > > +static const unsigned int mt8195_pull_type[] = {
> > > + MTK_PULL_PUPD_R1R0_TYPE /* 0 */, MTK_PULL_PUPD_R1R0_TYPE /*
> > > 1 */,
> > > + MTK_PULL_PUPD_R1R0_TYPE /* 2 */, MTK_PULL_PUPD_R1R0_TYPE /*
> > > 3 */,
> > > + MTK_PULL_PUPD_R1R0_TYPE /* 4 */, MTK_PULL_PUPD_R1R0_TYPE /*
> > > 5 */,
> > > + MTK_PULL_PU_PD_TYPE /* 6 */, MTK_PULL_PU_PD_TYPE /* 7 */,
> > > + MTK_PULL_PU_PD_RSEL_TYPE /* 8 */, MTK_PULL_PU_PD_RSEL_TYPE
> > > /* 9 */,
> > > + MTK_PULL_PU_PD_RSEL_TYPE /* 10 */, MTK_PULL_PU_PD_RSEL_TYPE
> > > /* 11 */,
> > > + MTK_PULL_PU_PD_RSEL_TYPE /* 12 */, MTK_PULL_PU_PD_RSEL_TYPE
> > > /* 13 */,
> > > + MTK_PULL_PU_PD_RSEL_TYPE /* 14 */, MTK_PULL_PU_PD_RSEL_TYPE
> > > /* 15 */,
> > > + MTK_PULL_PU_PD_RSEL_TYPE /* 16 */, MTK_PULL_PU_PD_RSEL_TYPE
> > > /* 17 */,
> > > + MTK_PULL_PU_PD_TYPE /* 18 */, MTK_PULL_PU_PD_TYPE /* 19 */,
> > > + MTK_PULL_PU_PD_TYPE /* 20 */, MTK_PULL_PU_PD_TYPE /* 21 */,
> > > + MTK_PULL_PU_PD_TYPE /* 22 */, MTK_PULL_PU_PD_TYPE /* 23 */,
> > > + MTK_PULL_PU_PD_TYPE /* 24 */, MTK_PULL_PU_PD_TYPE /* 25 */,
> > > + MTK_PULL_PU_PD_TYPE /* 26 */, MTK_PULL_PU_PD_TYPE /* 27 */,
> > > + MTK_PULL_PU_PD_TYPE /* 28 */, MTK_PULL_PU_PD_RSEL_TYPE /*
> > > 29 */,
> > > + MTK_PULL_PU_PD_RSEL_TYPE /* 30 */, MTK_PULL_PU_PD_TYPE /*
> > > 31 */,
> > > + MTK_PULL_PU_PD_TYPE /* 32 */, MTK_PULL_PU_PD_TYPE /* 33 */,
> > > + MTK_PULL_PU_PD_RSEL_TYPE /* 34 */, MTK_PULL_PU_PD_RSEL_TYPE
> > > /* 35 */,
> > > + MTK_PULL_PU_PD_TYPE /* 36 */, MTK_PULL_PU_PD_TYPE /* 37 */,
> > > + MTK_PULL_PU_PD_TYPE /* 38 */, MTK_PULL_PU_PD_TYPE /* 39 */,
> > > + MTK_PULL_PU_PD_TYPE /* 40 */, MTK_PULL_PU_PD_TYPE /* 41 */,
> > > + MTK_PULL_PU_PD_TYPE /* 42 */, MTK_PULL_PU_PD_TYPE /* 43 */,
> > > + MTK_PULL_PU_PD_RSEL_TYPE /* 44 */, MTK_PULL_PU_PD_RSEL_TYPE
> > > /* 45 */,
> > > + MTK_PULL_PU_PD_TYPE /* 46 */, MTK_PULL_PU_PD_TYPE /* 47 */,
> > > + MTK_PULL_PU_PD_TYPE /* 48 */, MTK_PULL_PU_PD_TYPE /* 49 */,
> > > + MTK_PULL_PU_PD_TYPE /* 50 */, MTK_PULL_PU_PD_TYPE /* 51 */,
> > > + MTK_PULL_PU_PD_TYPE /* 52 */, MTK_PULL_PU_PD_TYPE /* 53 */,
> > > + MTK_PULL_PU_PD_TYPE /* 54 */, MTK_PULL_PU_PD_TYPE /* 55 */,
> > > + MTK_PULL_PU_PD_TYPE /* 56 */, MTK_PULL_PU_PD_TYPE /* 57 */,
> > > + MTK_PULL_PU_PD_TYPE /* 58 */, MTK_PULL_PU_PD_TYPE /* 59 */,
> > > + MTK_PULL_PU_PD_TYPE /* 60 */, MTK_PULL_PU_PD_TYPE /* 61 */,
> > > + MTK_PULL_PU_PD_TYPE /* 62 */, MTK_PULL_PU_PD_TYPE /* 63 */,
> > > + MTK_PULL_PU_PD_TYPE /* 64 */, MTK_PULL_PU_PD_TYPE /* 65 */,
> > > + MTK_PULL_PU_PD_TYPE /* 66 */, MTK_PULL_PU_PD_TYPE /* 67 */,
> > > + MTK_PULL_PU_PD_TYPE /* 68 */, MTK_PULL_PU_PD_TYPE /* 69 */,
> > > + MTK_PULL_PU_PD_TYPE /* 70 */, MTK_PULL_PU_PD_TYPE /* 71 */,
> > > + MTK_PULL_PU_PD_TYPE /* 72 */, MTK_PULL_PU_PD_TYPE /* 73 */,
> > > + MTK_PULL_PU_PD_TYPE /* 74 */, MTK_PULL_PU_PD_TYPE /* 75 */,
> > > + MTK_PULL_PU_PD_TYPE /* 76 */, MTK_PULL_PUPD_R1R0_TYPE /* 77
> > > */,
> > > + MTK_PULL_PUPD_R1R0_TYPE /* 78 */, MTK_PULL_PUPD_R1R0_TYPE
> > > /* 79 */,
> > > + MTK_PULL_PUPD_R1R0_TYPE /* 80 */, MTK_PULL_PUPD_R1R0_TYPE
> > > /* 81 */,
> > > + MTK_PULL_PUPD_R1R0_TYPE /* 82 */, MTK_PULL_PUPD_R1R0_TYPE
> > > /* 83 */,
> > > + MTK_PULL_PUPD_R1R0_TYPE /* 84 */, MTK_PULL_PUPD_R1R0_TYPE
> > > /* 85 */,
> > > + MTK_PULL_PUPD_R1R0_TYPE /* 86 */, MTK_PULL_PUPD_R1R0_TYPE
> > > /* 87 */,
> > > + MTK_PULL_PUPD_R1R0_TYPE /* 88 */, MTK_PULL_PUPD_R1R0_TYPE
> > > /* 89 */,
> > > + MTK_PULL_PUPD_R1R0_TYPE /* 90 */, MTK_PULL_PUPD_R1R0_TYPE
> > > /* 91 */,
> > > + MTK_PULL_PUPD_R1R0_TYPE /* 92 */, MTK_PULL_PUPD_R1R0_TYPE
> > > /* 93 */,
> > > + MTK_PULL_PUPD_R1R0_TYPE /* 94 */, MTK_PULL_PUPD_R1R0_TYPE
> > > /* 95 */,
> > > + MTK_PULL_PUPD_R1R0_TYPE /* 96 */, MTK_PULL_PU_PD_TYPE /* 97
> > > */,
> > > + MTK_PULL_PU_PD_TYPE /* 98 */, MTK_PULL_PU_PD_TYPE /* 99 */,
> > > + MTK_PULL_PU_PD_TYPE /* 100 */, MTK_PULL_PU_PD_TYPE /* 101
> > > */,
> > > + MTK_PULL_PU_PD_TYPE /* 102 */, MTK_PULL_PU_PD_TYPE /* 103
> > > */,
> > > + MTK_PULL_PUPD_R1R0_TYPE /* 104 */, MTK_PULL_PUPD_R1R0_TYPE
> > > /* 105 */,
> > > + MTK_PULL_PUPD_R1R0_TYPE /* 106 */, MTK_PULL_PUPD_R1R0_TYPE
> > > /* 107 */,
> > > + MTK_PULL_PU_PD_TYPE /* 108 */, MTK_PULL_PU_PD_TYPE /* 109
> > > */,
> > > + MTK_PULL_PUPD_R1R0_TYPE /* 110 */, MTK_PULL_PUPD_R1R0_TYPE
> > > /* 111 */,
> > > + MTK_PULL_PUPD_R1R0_TYPE /* 112 */, MTK_PULL_PUPD_R1R0_TYPE
> > > /* 113 */,
> > > + MTK_PULL_PUPD_R1R0_TYPE /* 114 */, MTK_PULL_PUPD_R1R0_TYPE
> > > /* 115 */,
> > > + MTK_PULL_PUPD_R1R0_TYPE /* 116 */, MTK_PULL_PUPD_R1R0_TYPE
> > > /* 117 */,
> > > + MTK_PULL_PUPD_R1R0_TYPE /* 118 */, MTK_PULL_PUPD_R1R0_TYPE
> > > /* 119 */,
> > > + MTK_PULL_PUPD_R1R0_TYPE /* 120 */, MTK_PULL_PUPD_R1R0_TYPE
> > > /* 121 */,
> > > + MTK_PULL_PUPD_R1R0_TYPE /* 122 */, MTK_PULL_PUPD_R1R0_TYPE
> > > /* 123 */,
> > > + MTK_PULL_PUPD_R1R0_TYPE /* 124 */, MTK_PULL_PUPD_R1R0_TYPE
> > > /* 125 */,
> > > + MTK_PULL_PUPD_R1R0_TYPE /* 126 */, MTK_PULL_PUPD_R1R0_TYPE
> > > /* 127 */,
> > > + MTK_PULL_PU_PD_TYPE /* 128 */, MTK_PULL_PU_PD_TYPE /* 129
> > > */,
> > > + MTK_PULL_PU_PD_TYPE /* 130 */, MTK_PULL_PU_PD_TYPE /* 131
> > > */,
> > > + MTK_PULL_PU_PD_TYPE /* 132 */, MTK_PULL_PU_PD_TYPE /* 133
> > > */,
> > > + MTK_PULL_PU_PD_TYPE /* 134 */, MTK_PULL_PU_PD_TYPE /* 135
> > > */,
> > > + MTK_PULL_PU_PD_TYPE /* 136 */, MTK_PULL_PU_PD_TYPE /* 137
> > > */,
> > > + MTK_PULL_PU_PD_TYPE /* 138 */, MTK_PULL_PU_PD_TYPE /* 139
> > > */,
> > > + MTK_PULL_PU_PD_TYPE /* 140 */, MTK_PULL_PU_PD_TYPE /* 141
> > > */,
> > > + MTK_PULL_PU_PD_TYPE /* 142 */, MTK_PULL_PU_PD_TYPE /* 143
> > > */,
> > > +};
> > > +
> > > static const struct mtk_pin_reg_calc
> > > mt8195_reg_cals[PINCTRL_PIN_REG_MAX] = {
> > > [PINCTRL_PIN_REG_MODE] = MTK_RANGE(mt8195_pin_mode_range),
> > > [PINCTRL_PIN_REG_DIR] = MTK_RANGE(mt8195_pin_dir_range),
> > > @@ -793,6 +922,7 @@ static const struct mtk_pin_reg_calc
> > > mt8195_reg_cals[PINCTRL_PIN_REG_MAX] = {
> > > [PINCTRL_PIN_REG_R0] = MTK_RANGE(mt8195_pin_r0_range),
> > > [PINCTRL_PIN_REG_R1] = MTK_RANGE(mt8195_pin_r1_range),
> > > [PINCTRL_PIN_REG_DRV_ADV] =
> > > MTK_RANGE(mt8195_pin_drv_adv_range),
> > > + [PINCTRL_PIN_REG_RSEL] = MTK_RANGE(mt8195_pin_rsel_range),
> > > };
> > >
> > > static const char * const mt8195_pinctrl_register_base_names[] = {
> > > @@ -817,6 +947,9 @@ static const struct mtk_pin_soc mt8195_data = {
> > > .gpio_m = 0,
> > > .base_names = mt8195_pinctrl_register_base_names,
> > > .nbase_names =
> > > ARRAY_SIZE(mt8195_pinctrl_register_base_names),
> > > + .pull_type = mt8195_pull_type,
> > > + .pin_rsel = mt8195_pin_rsel_val_range,
> > > + .npin_rsel = ARRAY_SIZE(mt8195_pin_rsel_val_range),
> > > .bias_set_combo = mtk_pinconf_bias_set_combo,
> > > .bias_get_combo = mtk_pinconf_bias_get_combo,
> > > .drive_set = mtk_pinconf_drive_set_rev1,
> > > diff --git a/drivers/pinctrl/mediatek/pinctrl-mtk-common-v2.c
> > > b/drivers/pinctrl/mediatek/pinctrl-mtk-common-v2.c
> > > index 5b3b048725cc..b6a89d41f040 100644
> > > --- a/drivers/pinctrl/mediatek/pinctrl-mtk-common-v2.c
> > > +++ b/drivers/pinctrl/mediatek/pinctrl-mtk-common-v2.c
> > > @@ -641,6 +641,9 @@ static int
> > > mtk_pinconf_bias_set_pupd_r1_r0(struct mtk_pinctrl *hw,
> > > } else if (arg == MTK_PUPD_SET_R1R0_11) {
> > > r0 = 1;
> > > r1 = 1;
> > > + } else if (arg == MTK_ENABLE) {
> > > + r0 = 1;
> > > + r1 = 0;
> >
> > This change does not seem related. This should be in a separate patch
> > and why and what this change is should be described in the commit
> > log.
> > I assume this is to support "bias-pull-up" or "bias-pull-down"
> > without
> > arguments on pins using R1R0?
>
> we will separate it in next version.
> some users want only pull up/down, they don't care about the resistance
> value. So we support the case which only use "bias-pull-up" or "bias-
> pull-down" without arguments on pins using R1R0.

I see. This should be added as a separate patch before the RSEL stuff
then. That way it would also be easier to backport in case it is needed.

> >
> > > } else {
> > > err = -EINVAL;
> > > goto out;
> > > @@ -661,6 +664,75 @@ static int
> > > mtk_pinconf_bias_set_pupd_r1_r0(struct mtk_pinctrl *hw,
> > > return err;
> > > }
> > >
> > > +static int mtk_hw_pin_rsel_lookup(struct mtk_pinctrl *hw,
> > > + const struct mtk_pin_desc *desc,
> > > + u32 pullup, u32 arg, u32
> > > *rsel_val)
> > > +{
> > > + const struct mtk_pin_rsel *rsel;
> > > + int check;
> > > + bool found = false;
> > > +
> > > + rsel = hw->soc->pin_rsel;
> > > +
> > > + for (check = 0; check <= hw->soc->npin_rsel - 1; check++) {
> > > + if (desc->number >= rsel[check].s_pin &&
> > > + desc->number <= rsel[check].e_pin) {
> > > + if (pullup) {
> > > + if (rsel[check].up_rsel == arg) {
> > > + found = true;
> > > + *rsel_val =
> > > rsel[check].rsel_index;
> > > + break;
> > > + }
> > > + } else {
> > > + if (rsel[check].down_rsel == arg) {
> > > + found = true;
> > > + *rsel_val =
> > > rsel[check].rsel_index;
> > > + break;
> > > + }
> > > + }
> > > + }
> > > + }
> > > +
> > > + if (!found) {
> > > + dev_err(hw->dev, "Not support rsel value %d Ohm for
> > > pin = %d (%s)\n",
> > > + arg, desc->number, desc->name);
> > > + return -EOPNOTSUPP;
> >
> > I believe ENOTSUPP is the correct one. EOPNOTSUPP seems to be network
> > related.
> >
>
> if we change it as "ENOTSUPP", it will report warning"ENOTSUPP is not
> SUSV4 error code, prefer EOPNOTSUP" when checking patch.

The context surrounding this warning seems to be that ENOTSUPP is hard
for userspace to understand. AFAIK the return code here does not get
passed to userspace? And the pinctrl core does check for EINVAL or
ENOTSUPP, so I think this is a valid use case.

Linus?

> > I also think it would make sense to differentiate between the pin not
> > supporting RSEL (return -ENOTSUPP) vs supporting RSEL but a resistor
> > value not in the table was requested (return -EINVAL). This might be
> > easier to do with the two-table design I mentioned.
>
> It only support resistor value on the patch on mt8195. we think we need
> only maintain one-table design.

OK. I think there's a possibility that some corner case might happen.
I guess we'll deal with them if they occur.

> >
> > > + }
> > > +
> > > + return 0;
> > > +}
> > > +
> > > +static int mtk_pinconf_bias_set_rsel(struct mtk_pinctrl *hw,
> > > + const struct mtk_pin_desc
> > > *desc,
> > > + u32 pullup, u32 arg)
> > > +{
> > > + int err, rsel_val;
> > > +
> > > + if (hw->soc->pin_rsel) {
> > > + /* find pin rsel_index from pin_rsel array*/
> > > + err = mtk_hw_pin_rsel_lookup(hw, desc, pullup, arg,
> > > &rsel_val);
> > > + if (err)
> > > + goto out;
> > > + } else {
> > > + if (arg < MTK_PULL_SET_RSEL_000 ||
> > > + arg > MTK_PULL_SET_RSEL_111) {
> > > + err = -EINVAL;
> > > + goto out;
> > > + }
> > > +
> > > + rsel_val -= MTK_PULL_SET_RSEL_000;
> > > + }
> >
> > This is not structured correctly. Think about this: on a chip, say
> > MT8195,
> > which has hw->soc->pin_rsel defined, but the device tree writer chose
> > to use the macros instead of SI unit values. This is exactly the
> > scenario
> > you discussed privately with me.
> >
> > This also fails the case where bias-pull-{up,down} is specified
> > without
> > an argument, in which case arg == 1. This does seem to be covered by
> > falling back to just PU/PD, but it's not exactly obvious, especially
> > since we are skipping mtk_pinconf_bias_set_pu_pd() just below.
> >
> > So you actually need to check for the magic values first, and if arg
> > doesn't match any of them, assume an SI unit argument and try to do
> > a lookup.
> >
>
> one very important point of upstream is keeping backward compatibility,
> since almost all customers of other MTK platforms does not care about
> resistor values so we leave user to choose MTK_PULL_SET_RSEL_XXX
> without necessity to know resistor value.

I'm not sure about the connection to backward compatibility here, given
that you are just introducing support for RSEL.

> Our policy is making most customers/users convenient.
> Encourage of using SI unit is not in concern.
>
> For mt8195, we choose "hw->soc->pin_rsel", so it can support SI unit.

Regarding my concern, there are a lot of moving pieces here, so let's see
how the next revision turns out and if this is still an issue.

> > > +
> > > + err = mtk_hw_set_value(hw, desc, PINCTRL_PIN_REG_RSEL,
> > > rsel_val);
> > > + if (err)
> > > + goto out;
> > > +
> > > + err = mtk_pinconf_bias_set_pu_pd(hw, desc, pullup,
> > > MTK_ENABLE);
> > > +
> > > +out:
> > > + return err;
> > > +}
> > > +
> > > static int mtk_pinconf_bias_get_pu_pd(struct mtk_pinctrl *hw,
> > > const struct mtk_pin_desc *desc,
> > > u32 *pullup, u32 *enable)
> > > @@ -742,44 +814,117 @@ static int
> > > mtk_pinconf_bias_get_pupd_r1_r0(struct mtk_pinctrl *hw,
> >
> > This hunk is somewhat whacky. I assume it's because you also fixed up
> > the indentation for mtk_pinconf_bias_set_combo(). While nice to do,
> > in this case it actually works against having an easy to read patch.
> >
>
> yes, it's because to fix up
> the indentation for mtk_pinconf_bias_set_combo().

I suggest moving that to a separate patch. While whitespace changes are
sometimes frowned upon, I think this is OK given that it would improve
readability of the patches following it. You should mention that in the
commit message to justify the change.

> > > return err;
> > > }
> > >
> > > -int mtk_pinconf_bias_set_combo(struct mtk_pinctrl *hw,
> > > - const struct mtk_pin_desc *desc,
> > > - u32 pullup, u32 arg)
> > > +static int mtk_pinconf_bias_get_rsel(struct mtk_pinctrl *hw,
> > > + const struct mtk_pin_desc
> > > *desc,
> > > + u32 *pullup, u32 *enable)
> > > {
> > > - int err;
> > > + int pu, pd, rsel, err;
> > >
> > > - err = mtk_pinconf_bias_set_pu_pd(hw, desc, pullup, arg);
> > > - if (!err)
> > > + err = mtk_hw_get_value(hw, desc, PINCTRL_PIN_REG_RSEL,
> > > &rsel);
> > > + if (err)
> > > goto out;
> > >
> > > - err = mtk_pinconf_bias_set_pullsel_pullen(hw, desc, pullup,
> > > arg);
> > > - if (!err)
> > > + err = mtk_hw_get_value(hw, desc, PINCTRL_PIN_REG_PU, &pu);
> > > + if (err)
> > > goto out;
> > >
> > > - err = mtk_pinconf_bias_set_pupd_r1_r0(hw, desc, pullup,
> > > arg);
> > > + err = mtk_hw_get_value(hw, desc, PINCTRL_PIN_REG_PD, &pd);
> >
> > Maybe this could reuse mtk_pinconf_bias_get_pu_pd(), much like on the
> > `set`
> > side?
> >
>
> we will try to reuse mtk_pinconf_bias_get_pu_pd().
>
> > > +
> > > + if (pu == 0 && pd == 0) {
> > > + *pullup = 0;
> > > + *enable = MTK_DISABLE;
> > > + } else if (pu == 1 && pd == 0) {
> > > + *pullup = 1;
> > > + *enable = rsel + MTK_PULL_SET_RSEL_000;
> > > + } else if (pu == 0 && pd == 1) {
> > > + *pullup = 0;
> > > + *enable = rsel + MTK_PULL_SET_RSEL_000;
> > > + } else {
> > > + err = -EINVAL;
> > > + goto out;
> > > + }
> > >
> > > out:
> > > return err;
> > > }
> > > +
> > > +int mtk_pinconf_bias_set_combo(struct mtk_pinctrl *hw,
> > > + const struct mtk_pin_desc *desc,
> > > + u32 pullup, u32 arg)
> > > +{
> > > + int err = -EOPNOTSUPP;
> > > + bool try_all_type;
> > > +
> > > + try_all_type = hw->soc->pull_type ? false : true;
> >
> > Using ternary ops is discouraged. Maybe this could be structured
> > like:
> >
> > u32 pull_type;
> >
> > if (hw->soc->pull_type)
> > pull_type = hw->soc->pull_type[desc->number];
> > else
> > pull_type = MTK_PULL_TYPE_MASK;
> >
> > That way you don't need to test `try_all_type` every time.
>
> we will try to change it in next version.
>
> >
> > > +
> > > + if (try_all_type ||
> > > + (hw->soc->pull_type[desc->number] &
> > > MTK_PULL_RSEL_TYPE)) {
> > > + err = mtk_pinconf_bias_set_rsel(hw, desc, pullup,
> > > arg);
> > > + if (!err)
> > > + return err;
> > > + }
> > > +
> > > + if (try_all_type ||
> > > + (hw->soc->pull_type[desc->number] &
> > > MTK_PULL_PU_PD_TYPE)) {
> > > + err = mtk_pinconf_bias_set_pu_pd(hw, desc, pullup,
> > > arg);
> > > + if (!err)
> > > + return err;
> > > + }
> > > +
> > > + if (try_all_type ||
> > > + (hw->soc->pull_type[desc->number] &
> > > MTK_PULL_PULLSEL_TYPE)) {
> > > + err = mtk_pinconf_bias_set_pullsel_pullen(hw, desc,
> > > + pullup,
> > > arg);
> > > + if (!err)
> > > + return err;
> > > + }
> > > +
> > > + if (try_all_type ||
> > > + (hw->soc->pull_type[desc->number] &
> > > MTK_PULL_PUPD_R1R0_TYPE)) {
> > > + err = mtk_pinconf_bias_set_pupd_r1_r0(hw, desc,
> > > pullup, arg);
> > > + if (err)
> > > + dev_err(hw->dev, "Invalid pull
> > > argument\n");
> >
> > The "if (err)" could be moved outside the upper "if" block. That way
> > the code flow looks more consistent, plus we get an error message
> > regardless of the pull type supported.
>
> we will move "if (err)" outside the upper "if" block in the next
> version.
> >
> > > + }
> > > +
> > > + return err;
> > > +}
> > > EXPORT_SYMBOL_GPL(mtk_pinconf_bias_set_combo);
> > >
> > > int mtk_pinconf_bias_get_combo(struct mtk_pinctrl *hw,
> > > const struct mtk_pin_desc *desc,
> > > u32 *pullup, u32 *enable)
> > > {
> > > - int err;
> > > + int err = -EOPNOTSUPP;
> > > + bool try_all_type;
> > >
> > > - err = mtk_pinconf_bias_get_pu_pd(hw, desc, pullup, enable);
> > > - if (!err)
> > > - goto out;
> > > + try_all_type = hw->soc->pull_type ? false : true;
> >
> > Same here.
> >
>
> we will try to change it in next version.
>
> > >
> > > - err = mtk_pinconf_bias_get_pullsel_pullen(hw, desc, pullup,
> > > enable);
> > > - if (!err)
> > > - goto out;
> > > + if (try_all_type ||
> > > + (hw->soc->pull_type[desc->number] &
> > > MTK_PULL_RSEL_TYPE)) {
> > > + err = mtk_pinconf_bias_get_rsel(hw, desc, pullup,
> > > enable);
> > > + if (!err)
> > > + return err;
> > > + }
> > > +
> > > + if (try_all_type ||
> > > + (hw->soc->pull_type[desc->number] &
> > > MTK_PULL_PU_PD_TYPE)) {
> > > + err = mtk_pinconf_bias_get_pu_pd(hw, desc, pullup,
> > > enable);
> > > + if (!err)
> > > + return err;
> > > + }
> > >
> > > - err = mtk_pinconf_bias_get_pupd_r1_r0(hw, desc, pullup,
> > > enable);
> > > + if (try_all_type ||
> > > + (hw->soc->pull_type[desc->number] &
> > > MTK_PULL_PULLSEL_TYPE)) {
> > > + err = mtk_pinconf_bias_get_pullsel_pullen(hw, desc,
> > > + pullup,
> > > enable);
> > > + if (!err)
> > > + return err;
> > > + }
> > > +
> > > + if (try_all_type ||
> > > + (hw->soc->pull_type[desc->number] &
> > > MTK_PULL_PUPD_R1R0_TYPE))
> > > + err = mtk_pinconf_bias_get_pupd_r1_r0(hw, desc,
> > > pullup, enable);
> > >
> > > -out:
> > > return err;
> > > }
> > > EXPORT_SYMBOL_GPL(mtk_pinconf_bias_get_combo);
> > > diff --git a/drivers/pinctrl/mediatek/pinctrl-mtk-common-v2.h
> > > b/drivers/pinctrl/mediatek/pinctrl-mtk-common-v2.h
> > > index a6f1bdb2083b..a3d73f153efe 100644
> > > --- a/drivers/pinctrl/mediatek/pinctrl-mtk-common-v2.h
> > > +++ b/drivers/pinctrl/mediatek/pinctrl-mtk-common-v2.h
> > > @@ -17,6 +17,13 @@
> > > #define MTK_ENABLE 1
> > > #define MTK_PULLDOWN 0
> > > #define MTK_PULLUP 1
> > > +#define MTK_PULL_PU_PD_TYPE BIT(0)
> > > +#define MTK_PULL_PULLSEL_TYPE BIT(1)
> > > +#define MTK_PULL_PUPD_R1R0_TYPE BIT(2)
> > > +#define MTK_PULL_RSEL_TYPE BIT(3)
> > > +#define MTK_PULL_PU_PD_RSEL_TYPE (MTK_PULL_PU_PD_TYPE \
> > > + | MTK_PULL_RSEL_TYPE)
> >
> > MTK_PULL_RSEL_TYPE deserves a comment saying why it should not be
> > used
> > on its own.
> >
> > But since mtk_pinconf_bias_set_rsel() is already calling
> > mtk_pinconf_bias_set_pu_pd(), if the former function is reworked,
> > then
> > the two types don't need to be combined for the fallback behavior.
> >
>
>
> MTK_PULL_PU_PD_RSEL_TYPE is type wihch is can be control by
> MTK_PULL_PU_PD_TYPE or MTK_PULL_RSEL_TYPE. so we combibe them
> together.

But can RSEL be used without PU/PD? The definitions read like "PU_PD_RSEL"
means either RSEL or PU/PD can be used, but if I understand the hardware
correctly, RSEL by itself only selects the resistor value, but doesn't
turn on/off the bias nor select pull up or down. So RSEL is an extension
of PU/PD, and by itself won't have any real effect.

Since this could be interpreted either way, a comment explaining things
should be added to document the author's intentions.

> > > #define EINT_NA U16_MAX
> > > #define NO_EINT_SUPPORT EINT_NA
> > > @@ -42,6 +49,14 @@
> > > PIN_FIELD_CALC(_s_pin, _e_pin, 0, _s_addr, _x_addrs,
> > > _s_bit, \
> > > _x_bits, 32, 1)
> > >
> > > +#define PIN_RSEL(_s_pin, _e_pin, _rsel_index, _up_resl,
> > > _down_rsel) { \
> >
> > ^ rsel?
> >
> > > + .s_pin =
> > > _s_pin, \
> > > + .e_pin =
> > > _e_pin, \
> > > + .rsel_index =
> > > _rsel_index, \
> > > + .up_rsel =
> > > _up_resl, \
> > > + .down_rsel =
> > > _down_rsel, \
> > > + }
> > > +
> > > /* List these attributes which could be modified for the pin */
> > > enum {
> > > PINCTRL_PIN_REG_MODE,
> > > @@ -67,6 +82,7 @@ enum {
> > > PINCTRL_PIN_REG_DRV_E0,
> > > PINCTRL_PIN_REG_DRV_E1,
> > > PINCTRL_PIN_REG_DRV_ADV,
> > > + PINCTRL_PIN_REG_RSEL,
> > > PINCTRL_PIN_REG_MAX,
> > > };
> > >
> > > @@ -129,6 +145,21 @@ struct mtk_pin_field_calc {
> > > u8 fixed;
> > > };
> > >
> > > +/* struct mtk_pin_rsel - the structure that providing bias
> > > resistance selection.
> >
> > ^ provides
>
> we will change it in next version.
>
>
> >
> > > + * @s_pin: the start pin within the rsel range
> > > + * @e_pin: the end pin within the rsel range
> > > + * @rsel_index: the rsel bias resistance index
> > > + * @up_rsel: the pullup rsel bias resistance value
> > > + * @down_rsel: the pulldown rsel bias resistance value
> > > + */
> > > +struct mtk_pin_rsel {
> > > + u16 s_pin;
> > > + u16 e_pin;
> > > + u16 rsel_index;
> > > + u32 up_rsel;
> > > + u32 down_rsel;
> > > +};
> > > +
> >
> > If it were up to me, I would split this into two layers, one for the
> > register value <-> resistor value mapping, lets call it "rsel table",
> > and the other for pin <-> "rsel table" mapping.
> >
> > I assume most if not all pins that support rsel would have the same
> > set
> > of resistor values, so that would trade a level of indirection for
> > better
> > usage of space.
> >
> > > /* struct mtk_pin_reg_calc - the structure that holds all ranges
> > > used to
> > > * determine which register the pin would
> > > make use of
> > > * for certain pin attribute.
> > > @@ -206,6 +237,9 @@ struct mtk_pin_soc {
> > > bool ies_present;
> > > const char * const *base_names;
> > > unsigned int nbase_names;
> > > + const unsigned int *pull_type;
> > > + const struct mtk_pin_rsel *pin_rsel;
> > > + unsigned int npin_rsel;
> > >
> > > /* Specific pinconfig operations */
> > > int (*bias_disable_set)(struct mtk_pinctrl *hw,
> > > @@ -237,7 +271,6 @@ struct mtk_pin_soc {
> > > const struct mtk_pin_desc *desc, u32
> > > arg);
> > > int (*adv_drive_get)(struct mtk_pinctrl *hw,
> > > const struct mtk_pin_desc *desc, u32
> > > *val);
> > > -
> >
> > Unrelated whitespace change. Please remove it.
>
> we will remove it in next version.
>
> >
> > > /* Specific driver data */
> > > void *driver_data;
> > > };
> > > diff --git a/drivers/pinctrl/mediatek/pinctrl-paris.c
> > > b/drivers/pinctrl/mediatek/pinctrl-paris.c
> > > index 85db2e4377f0..8990cfe47d72 100644
> > > --- a/drivers/pinctrl/mediatek/pinctrl-paris.c
> > > +++ b/drivers/pinctrl/mediatek/pinctrl-paris.c
> > > @@ -577,9 +577,9 @@ static int mtk_hw_get_value_wrap(struct
> > > mtk_pinctrl *hw, unsigned int gpio, int
> > > mtk_hw_get_value_wrap(hw, gpio, PINCTRL_PIN_REG_DRV)
> > >
> > > ssize_t mtk_pctrl_show_one_pin(struct mtk_pinctrl *hw,
> > > - unsigned int gpio, char *buf, unsigned int bufLen)
> > > + unsigned int gpio, char *buf, unsigned int buf_len)
> >
> > Unrelated change.
>
> we will sparate it in next version.
>
> >
> > > {
> > > - int pinmux, pullup, pullen, len = 0, r1 = -1, r0 = -1;
> > > + int pinmux, pullup, pullen, len = 0, r1 = -1, r0 = -1, rsel
> > > = -1;
> > > const struct mtk_pin_desc *desc;
> > >
> > > if (gpio >= hw->soc->npins)
> > > @@ -591,6 +591,8 @@ ssize_t mtk_pctrl_show_one_pin(struct
> > > mtk_pinctrl *hw,
> > > pinmux -= hw->soc->nfuncs;
> > >
> > > mtk_pinconf_bias_get_combo(hw, desc, &pullup, &pullen);
> > > +
> > > + /* Case for: R1R0 */
> > > if (pullen == MTK_PUPD_SET_R1R0_00) {
> > > pullen = 0;
> > > r1 = 0;
> > > @@ -607,10 +609,16 @@ ssize_t mtk_pctrl_show_one_pin(struct
> > > mtk_pinctrl *hw,
> > > pullen = 1;
> > > r1 = 1;
> > > r0 = 1;
> > > - } else if (pullen != MTK_DISABLE && pullen != MTK_ENABLE) {
> > > - pullen = 0;
> > > }
> > > - len += scnprintf(buf + len, bufLen - len,
> > > +
> > > + /* Case for: RSEL */
> > > + if (pullen >= MTK_PULL_SET_RSEL_000 &&
> > > + pullen <= MTK_PULL_SET_RSEL_111) {
> > > + rsel = pullen - MTK_PULL_SET_RSEL_000;
> > > + pullen = 1;
> > > + }
> > > +
> > > + len += scnprintf(buf + len, buf_len - len,
> >
> > Unrelated change.
>
> it is used to get rsel debug changes
> >
> > > "%03d: %1d%1d%1d%1d%02d%1d%1d%1d%1d",
> > > gpio,
> > > pinmux,
> > > @@ -624,10 +632,12 @@ ssize_t mtk_pctrl_show_one_pin(struct
> > > mtk_pinctrl *hw,
> > > pullup);
> > >
> > > if (r1 != -1) {
> > > - len += scnprintf(buf + len, bufLen - len, " (%1d
> > > %1d)\n",
> > > + len += scnprintf(buf + len, buf_len - len, " (%1d
> > > %1d)\n",
> > > r1, r0);
> > > + } else if (rsel != -1) {
> > > + len += scnprintf(buf + len, buf_len - len, "
> > > (%1d)\n", rsel);
> > > } else {
> > > - len += scnprintf(buf + len, bufLen - len, "\n");
> > > + len += scnprintf(buf + len, buf_len - len, "\n");
> >
> > Unrelated changes.
>
> it is used to get rsel debug changes

Sorry about the confusion. The unrelated change I meant to point out
is actually just the "bufLen" to "buf_len" change.

Regards
ChenYu

> >
> > > }
> > >
> > > return len;
> > > diff --git a/drivers/pinctrl/mediatek/pinctrl-paris.h
> > > b/drivers/pinctrl/mediatek/pinctrl-paris.h
> > > index afb7650fd25b..681267c0e1a4 100644
> > > --- a/drivers/pinctrl/mediatek/pinctrl-paris.h
> > > +++ b/drivers/pinctrl/mediatek/pinctrl-paris.h
> > > @@ -61,7 +61,7 @@ int mtk_paris_pinctrl_probe(struct
> > > platform_device *pdev,
> > > const struct mtk_pin_soc *soc);
> > >
> > > ssize_t mtk_pctrl_show_one_pin(struct mtk_pinctrl *hw,
> > > - unsigned int gpio, char *buf, unsigned int bufLen);
> > > + unsigned int gpio, char *buf, unsigned int buf_len);
> >
> > Unrelated change.
> >
>
> sparate it in next version.
>
> >
> > Regards
> > ChenYu
> >
> > >
> > > extern const struct dev_pm_ops mtk_paris_pinctrl_pm_ops;
> > >
> > > --
> > > 2.18.0
> > > _______________________________________________
> > > Linux-mediatek mailing list
> > > Linux-mediatek@xxxxxxxxxxxxxxxxxxx
> > > http://lists.infradead.org/mailman/listinfo/linux-mediatekwe will
>
>
>

Next message: Ryan Chen: "RE: [PATCH v3] media: aspeed-video: ignore interrupts that aren't enabled"
Previous message: Paolo Bonzini: "Re: [PATCH 1/2] KVM: VMX: avoid running vmx_handle_exit_irqoff in case of emulation"
In reply to: Chen-Yu Tsai: "Re: [PATCH v11 4/4] pinctrl: mediatek: add rsel setting on MT8195"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]