Re: [PATCH 5/6] drivers/thermal/exynos: add initial Exynos 850 support

From: Sam Protsenko
Date: Wed Jul 24 2024 - 20:56:57 EST


On Wed, Jul 24, 2024 at 10:31 AM Mateusz Majewski
<m.majewski2@xxxxxxxxxxx> wrote:
>
> > I'd suggest to group all the definitions here as such:
> >
> > #define REG1_OFFSET
> > #define REG1_FIELD1_OFFSET
> > #define REG1_FIELD2_OFFSET
> > ...empty line...
> > #define REG2_OFFSET
> > #define REG2_FIELD1_OFFSET
> > #define REG2_FIELD2_OFFSET
> > ...etc...
> >
> > Or otherwise each shift/mask constant should contain its register name
> > as a prefix, to avoid confusion. But right now it's kinda hard to
> > understand what belongs to what :) But that's just a nitpick.
>
> I came up with this:
>
> /* Exynos850 specific registers */
> #define EXYNOS850_TMU_REG_CURRENT_TEMP0_1 0x40
> #define EXYNOS850_TMU_REG_THD_TEMP0_RISE 0x50
> #define EXYNOS850_TMU_REG_THD_TEMP0_FALL 0x60
> #define EXYNOS850_TMU_TEMP_SHIFT 9
>
> #define EXYNOS850_TMU_TRIMINFO_SHIFT 4
> #define EXYNOS850_TMU_TRIMINFO_OFFSET(n) \
> (EXYNOS_TMU_REG_TRIMINFO + (n) * EXYNOS850_TMU_TRIMINFO_SHIFT)
> #define EXYNOS850_TMU_T_TRIM0_SHIFT 18
>
> #define EXYNOS850_TMU_REG_CONTROL1 0x24
> #define EXYNOS850_TMU_LPI_MODE_MASK 1
> #define EXYNOS850_TMU_LPI_MODE_SHIFT 10
>
> #define EXYNOS850_TMU_REG_COUNTER_VALUE0 0x30
> #define EXYNOS850_TMU_EN_TEMP_SEN_OFF_MASK 0xffff
> #define EXYNOS850_TMU_EN_TEMP_SEN_OFF_SHIFT 0
>
> #define EXYNOS850_TMU_REG_COUNTER_VALUE1 0x34
> #define EXYNOS850_TMU_CLK_SENSE_ON_MASK 0xffff
> #define EXYNOS850_TMU_CLK_SENSE_ON_SHIFT 16
>
> #define EXYNOS850_TMU_REG_AVG_CON 0x38
> #define EXYNOS850_TMU_AVG_MODE_MASK 0x7
> #define EXYNOS850_TMU_DEM_ENABLE BIT(4)
>
> #define EXYNOS850_TMU_REG_TRIM0 0x3c
> #define EXYNOS850_TMU_TRIM0_MASK 0xf
> #define EXYNOS850_TMU_VBEI_TRIM_SHIFT 8
> #define EXYNOS850_TMU_VREF_TRIM_SHIFT 12
> #define EXYNOS850_TMU_BGRI_TRIM_SHIFT 20
>
> #define EXYNOS850_TMU_TEM1051X_SENSE_VALUE 0x028a
> #define EXYNOS850_TMU_TEM1456X_SENSE_VALUE 0x0a28
>

Looks better, thanks!

> This also omits some definitions that were in v1, as they had the same
> value and they were the same thing anyway. For instance, I dropped
> EXYNOS850_TMU_T_BUF_VREF_SEL_MASK in favor of
> EXYNOS_TMU_REF_VOLTAGE_MASK, and have a single EXYNOS850_TMU_TRIM0_MASK
> instead of EXYNOS850_TMU_BGRI_TRIM_MASK, EXYNOS850_TMU_VREF_TRIM_MASK,
> EXYNOS850_TMU_VBEI_TRIM_MASK and EXYNOS850_TMU_T_TRIM0_MASK. Also,
>
> > Suggest using GENMASK() macro whenever possible.
>
> This would make me have a separate mask for each of these again. Maybe
> if this driver gets refactored in the future to use u32_get_bits() and
> so on this would make more sense?

Sure, that was just a suggestion, don't have a strong opinion on that
one. If you don't like it, feel free to skip it for now.