| Message ID | 20220830171155.42962-1-andriy.shevchenko@linux.intel.com |
|---|---|
| State | Superseded |
| Headers | show |
| Series | [v2,1/3] ACPI: PMIC: Use sizeof() instead of hard coded value | expand |
On Tue, Aug 30, 2022 at 08:11:54PM +0300, Andy Shevchenko wrote: > -#define VR_MODE_DISABLED 0 > -#define VR_MODE_AUTO BIT(0) > -#define VR_MODE_NORMAL BIT(1) > -#define VR_MODE_SWITCH BIT(2) > -#define VR_MODE_ECO (BIT(0)|BIT(1)) > +#define PMIC_REG_MASK GENMASK(11, 0) > + > +#define VR_MODE_DISABLED (0 << 0) > +#define VR_MODE_AUTO (1 << 0) > +#define VR_MODE_NORMAL (2 << 0) > +#define VR_MODE_ECO (3 << 0) > +#define VR_MODE_SWITCH (4 << 0) IMHO this one is worse than what it was. Anyway, that's just a nitpick. The other parts look good, Reviewed-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Hi, On 8/31/22 11:34, Andy Shevchenko wrote: > On Wed, Aug 31, 2022 at 08:43:54AM +0300, Mika Westerberg wrote: >> On Tue, Aug 30, 2022 at 08:11:54PM +0300, Andy Shevchenko wrote: >>> -#define VR_MODE_DISABLED 0 >>> -#define VR_MODE_AUTO BIT(0) >>> -#define VR_MODE_NORMAL BIT(1) >>> -#define VR_MODE_SWITCH BIT(2) >>> -#define VR_MODE_ECO (BIT(0)|BIT(1)) >>> +#define PMIC_REG_MASK GENMASK(11, 0) >>> + >>> +#define VR_MODE_DISABLED (0 << 0) >>> +#define VR_MODE_AUTO (1 << 0) >>> +#define VR_MODE_NORMAL (2 << 0) >>> +#define VR_MODE_ECO (3 << 0) >>> +#define VR_MODE_SWITCH (4 << 0) >> >> IMHO this one is worse than what it was. > > I'm not sure why. Here is obvious wrong use of BIT() macro against > plain numbers. I can split it into a separate change with an explanation > of why it's better. But I think it doesn't worth the churn. FWIW I'm with Andy here, the VR_MODE_ECO clearly is trying to just say 3, so this is just a plain enum for values 0-4 and as such should not use the BIT macros. Regards, Hans >> Anyway, that's just a nitpick. The other parts look good, >> >> Reviewed-by: Mika Westerberg <mika.westerberg@linux.intel.com> > > Thanks! >
From: Mika Westerberg > Sent: 31 August 2022 10:49 > > On Wed, Aug 31, 2022 at 11:37:21AM +0200, Hans de Goede wrote: > > Hi, > > > > On 8/31/22 11:34, Andy Shevchenko wrote: > > > On Wed, Aug 31, 2022 at 08:43:54AM +0300, Mika Westerberg wrote: > > >> On Tue, Aug 30, 2022 at 08:11:54PM +0300, Andy Shevchenko wrote: > > >>> -#define VR_MODE_DISABLED 0 > > >>> -#define VR_MODE_AUTO BIT(0) > > >>> -#define VR_MODE_NORMAL BIT(1) > > >>> -#define VR_MODE_SWITCH BIT(2) > > >>> -#define VR_MODE_ECO (BIT(0)|BIT(1)) > > >>> +#define PMIC_REG_MASK GENMASK(11, 0) > > >>> + > > >>> +#define VR_MODE_DISABLED (0 << 0) > > >>> +#define VR_MODE_AUTO (1 << 0) > > >>> +#define VR_MODE_NORMAL (2 << 0) > > >>> +#define VR_MODE_ECO (3 << 0) > > >>> +#define VR_MODE_SWITCH (4 << 0) > > >> > > >> IMHO this one is worse than what it was. > > > > > > I'm not sure why. Here is obvious wrong use of BIT() macro against > > > plain numbers. I can split it into a separate change with an explanation > > > of why it's better. But I think it doesn't worth the churn. > > > > FWIW I'm with Andy here, the VR_MODE_ECO clearly is trying > > to just say 3, so this is just a plain enum for values 0-4 and > > as such should not use the BIT macros. > > Yeah, enum would look better but the << 0 just makes me confused ;-) No idea what that code is doing. The values are all used to initialise a .bit structure member. So maybe BIT() is right. The _ECO value isn't used at all. Deeper analysis may be needed. David - Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK Registration No: 1397386 (Wales)
On Wed, Aug 31, 2022 at 10:06:09AM +0000, David Laight wrote: > From: Mika Westerberg > > Sent: 31 August 2022 10:49 > > On Wed, Aug 31, 2022 at 11:37:21AM +0200, Hans de Goede wrote: > > > On 8/31/22 11:34, Andy Shevchenko wrote: > > > > On Wed, Aug 31, 2022 at 08:43:54AM +0300, Mika Westerberg wrote: > > > >> On Tue, Aug 30, 2022 at 08:11:54PM +0300, Andy Shevchenko wrote: > > > >>> -#define VR_MODE_DISABLED 0 > > > >>> -#define VR_MODE_AUTO BIT(0) > > > >>> -#define VR_MODE_NORMAL BIT(1) > > > >>> -#define VR_MODE_SWITCH BIT(2) > > > >>> -#define VR_MODE_ECO (BIT(0)|BIT(1)) > > > >>> +#define PMIC_REG_MASK GENMASK(11, 0) > > > >>> + > > > >>> +#define VR_MODE_DISABLED (0 << 0) > > > >>> +#define VR_MODE_AUTO (1 << 0) > > > >>> +#define VR_MODE_NORMAL (2 << 0) > > > >>> +#define VR_MODE_ECO (3 << 0) > > > >>> +#define VR_MODE_SWITCH (4 << 0) > > > >> > > > >> IMHO this one is worse than what it was. > > > > > > > > I'm not sure why. Here is obvious wrong use of BIT() macro against > > > > plain numbers. I can split it into a separate change with an explanation > > > > of why it's better. But I think it doesn't worth the churn. > > > > > > FWIW I'm with Andy here, the VR_MODE_ECO clearly is trying > > > to just say 3, so this is just a plain enum for values 0-4 and > > > as such should not use the BIT macros. > > > > Yeah, enum would look better but the << 0 just makes me confused ;-) > > No idea what that code is doing. > The values are all used to initialise a .bit structure member. > So maybe BIT() is right. > The _ECO value isn't used at all. > > Deeper analysis may be needed. So, can you do that since you already started?
diff --git a/drivers/acpi/pmic/intel_pmic_chtdc_ti.c b/drivers/acpi/pmic/intel_pmic_chtdc_ti.c index 418eec523025..6c2a6da430ed 100644 --- a/drivers/acpi/pmic/intel_pmic_chtdc_ti.c +++ b/drivers/acpi/pmic/intel_pmic_chtdc_ti.c @@ -87,7 +87,7 @@ static int chtdc_ti_pmic_get_raw_temp(struct regmap *regmap, int reg) { u8 buf[2]; - if (regmap_bulk_read(regmap, reg, buf, 2)) + if (regmap_bulk_read(regmap, reg, buf, sizeof(buf))) return -EIO; /* stored in big-endian */ diff --git a/drivers/acpi/pmic/intel_pmic_xpower.c b/drivers/acpi/pmic/intel_pmic_xpower.c index 61bbe4c24d87..33c5e85294cd 100644 --- a/drivers/acpi/pmic/intel_pmic_xpower.c +++ b/drivers/acpi/pmic/intel_pmic_xpower.c @@ -255,7 +255,7 @@ static int intel_xpower_pmic_get_raw_temp(struct regmap *regmap, int reg) if (ret) return ret; - ret = regmap_bulk_read(regmap, AXP288_GP_ADC_H, buf, 2); + ret = regmap_bulk_read(regmap, AXP288_GP_ADC_H, buf, sizeof(buf)); if (ret == 0) ret = (buf[0] << 4) + ((buf[1] >> 4) & 0x0f);
It's better to use sizeof() of a given buffer than spreading a hard coded value. Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> --- v2: updated another driver as well (due to this no tag added) drivers/acpi/pmic/intel_pmic_chtdc_ti.c | 2 +- drivers/acpi/pmic/intel_pmic_xpower.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)