iio: light: si1133: driver cleanup
v2 cover iio: light: si1133: driver cleanup
This series deals with the cleanup and modernization of the Silicon Labs SI1133 UV/Ambient light sensor.
What seemed like a simple checkpatch cleanup turned out to be more complex, therefore I've ended up doing a patch series instead.
Changes include:
- adding missing headers
- removing unused macros
- adding guard(mutex)() support
- code style fixes
---
Changes in v2:
- Change v2 to patch series
- PATCH 2: remove unused macros
- PATCH 3: add missing include headers
- PATCH 4: group generic <linux/*> headers
- PATCH 5: add guard(mutex)() support
- Link to v1: https://lore.kernel.org/r/[email protected]
---
Joshua Crofts (5):
iio: light: si1133: prefer complex macros enclosed in parenthesis
iio: light: si1133: remove unused macros
iio: light: si1133: add missing include headers
iio: light: si1133: group generic <linux/*> headers
iio: light: si1133: use guard(mutex)() macro
drivers/iio/light/si1133.c | 57 ++++++++++++++++++++++------------------------ 1 file changed, 27 insertions(+), 30 deletions(-)
---
base-commit: d86db1905add39f905cf9f04252804b359914ed6
change-id: 20260427-si1133-checkup-abcc5280adf3
Best regards,
Signature · 1 line
Joshua Crofts <[email protected]>
PATCH v2 1/5 iio: light: si1133: prefer complex macros enclosed in parenthesis
Enclose complex macros in parenthesis per checkpatch.pl error to improve code style.
--- drivers/iio/light/si1133.c | 18 +++++++++--------- 1 file changed, 9 insertions(+), 9 deletions(-)
drivers/iio/light/si1133.c +9 -9
@@ -50,23 +50,23 @@#define SI1133_MAX_CMD_CTR 0xF#define SI1133_PARAM_REG_CHAN_LIST 0x01 -#define SI1133_PARAM_REG_ADCCONFIG (x) ((x) * 4) + 2 -#define SI1133_PARAM_REG_ADCSENS (x) ((x) * 4) + 3 -#define SI1133_PARAM_REG_ADCPOST (x) ((x) * 4) + 4 +#define SI1133_PARAM_REG_ADCCONFIG (x) (((x) * 4) + 2) +#define SI1133_PARAM_REG_ADCSENS (x) (((x) * 4) + 3) +#define SI1133_PARAM_REG_ADCPOST (x) (((x) * 4) + 4)#define SI1133_ADCMUX_MASK 0x1F -#define SI1133_ADCCONFIG_DECIM_RATE (x) (x) << 5 +#define SI1133_ADCCONFIG_DECIM_RATE (x) ((x) << 5)#define SI1133_ADCSENS_SCALE_MASK 0x70#define SI1133_ADCSENS_SCALE_SHIFT 4#define SI1133_ADCSENS_HSIG_MASK BIT(7)#define SI1133_ADCSENS_HSIG_SHIFT 7#define SI1133_ADCSENS_HW_GAIN_MASK 0xF -#define SI1133_ADCSENS_NB_MEAS (x) fls(x) << SI1133_ADCSENS_SCALE_SHIFT +#define SI1133_ADCSENS_NB_MEAS (x) (fls(x) << SI1133_ADCSENS_SCALE_SHIFT)#define SI1133_ADCPOST_24BIT_EN BIT(6) -#define SI1133_ADCPOST_POSTSHIFT_BITQTY (x) (x & GENMASK(2, 0)) << 3 +#define SI1133_ADCPOST_POSTSHIFT_BITQTY (x) (((x) & GENMASK(2, 0)) << 3)#define SI1133_PARAM_ADCMUX_SMALL_IR 0x0#define SI1133_PARAM_ADCMUX_MED_IR 0x1
@@ -87,15 +87,15 @@#define SI1133_CMD_MINSLEEP_US_HIGH 7500#define SI1133_CMD_TIMEOUT_MS 25#define SI1133_CMD_LUX_TIMEOUT_MS 5000 -#define SI1133_CMD_TIMEOUT_US SI1133_CMD_TIMEOUT_MS * 1000 +#define SI1133_CMD_TIMEOUT_US (SI1133_CMD_TIMEOUT_MS * 1000) -#define SI1133_REG_HOSTOUT (x) (x) + 0x13 +#define SI1133_REG_HOSTOUT (x) ((x) + 0x13)#define SI1133_MEASUREMENT_FREQUENCY 1250#define SI1133_X_ORDER_MASK 0x0070#define SI1133_Y_ORDER_MASK 0x0007 -#define si1133_get_x_order (m) ((m) & SI1133_X_ORDER_MASK) >> 4 +#define si1133_get_x_order (m) (((m) & SI1133_X_ORDER_MASK) >> 4)#define si1133_get_y_order (m) ((m) & SI1133_Y_ORDER_MASK)#define SI1133_LUX_ADC_MASK 0xE
Signature · 1 line
2.47.3
PATCH v2 2/5 iio: light: si1133: remove unused macros
In reply to Joshua CroftsRemove unused macros to improve readability.
No functional change.
--- drivers/iio/light/si1133.c | 7 ------- 1 file changed, 7 deletions(-)
drivers/iio/light/si1133.c +0 -7
@@ -22,8 +22,6 @@#define SI1133_REG_PART_ID 0x00#define SI1133_REG_REV_ID 0x01#define SI1133_REG_MFR_ID 0x02 -#define SI1133_REG_INFO0 0x03 -#define SI1133_REG_INFO1 0x04#define SI1133_PART_ID 0x33
@@ -40,7 +38,6 @@#define SI1133_CMD_RESET_CTR 0x00#define SI1133_CMD_RESET_SW 0x01#define SI1133_CMD_FORCE 0x11 -#define SI1133_CMD_START_AUTONOMOUS 0x13#define SI1133_CMD_PARAM_SET 0x80#define SI1133_CMD_PARAM_QUERY 0x40#define SI1133_CMD_PARAM_MASK 0x3F
@@ -86,13 +83,9 @@#define SI1133_CMD_MINSLEEP_US_LOW 5000#define SI1133_CMD_MINSLEEP_US_HIGH 7500#define SI1133_CMD_TIMEOUT_MS 25 -#define SI1133_CMD_LUX_TIMEOUT_MS 5000 -#define SI1133_CMD_TIMEOUT_US (SI1133_CMD_TIMEOUT_MS * 1000)#define SI1133_REG_HOSTOUT (x) ((x) + 0x13) -#define SI1133_MEASUREMENT_FREQUENCY 1250 -#define SI1133_X_ORDER_MASK 0x0070#define SI1133_Y_ORDER_MASK 0x0007#define si1133_get_x_order (m) (((m) & SI1133_X_ORDER_MASK) >> 4)
Signature · 1 line
2.47.3
Jonathan Cameron
maintainerRe: [PATCH v2 2/5] iio: light: si1133: remove unused macros
Quoted 3 lines · On Tue, 28 Apr 2026 15:18:34 +0200 Joshua Crofts via B4 Relay…
From: Joshua Crofts <[email protected]>
Remove unused macros to improve readability.
Hi Joshua,
Don't update macros then remove them - that's just increasing noise in the update patch. Reorder this patch to be before patch 1.
Otherwise, I think a few of these can stay. See below
Thanks,
Jonathan
Quoted 13 lines
No functional change.
Signed-off-by Joshua Crofts <[email protected]>--- drivers/iio/light/si1133.c | 7 ------- 1 file changed, 7 deletions(-)drivers/iio/light/si1133.c +0
-0
drivers/iio/light/si1133.c line 26
Quoted 5 lines
drivers/iio/light/si1133.c +0
-2@@ -22,8 +22,6 @@#define SI1133_REG_PART_ID 0x00#define SI1133_REG_REV_ID 0x01#define SI1133_REG_MFR_ID 0x02 -#define SI1133_REG_INFO0 0x03 -#define SI1133_REG_INFO1 0x04
This one isn't in the category of things Andy was referring to.
It's a hardware register address definition. We don't have hard rules on whether a complete set of these should be defined, or just the ones used but it is definitely acceptable to just define them all up front as then easy to check them against a datasheet.
drivers/iio/light/si1133.c line 43
Quoted 8 lines
drivers/iio/light/si1133.c +0
-1#define SI1133_PART_ID 0x33@@ -40,7 +38,6 @@#define SI1133_CMD_RESET_CTR 0x00#define SI1133_CMD_RESET_SW 0x01#define SI1133_CMD_FORCE 0x11 -#define SI1133_CMD_START_AUTONOMOUS 0x13
Similar for this.
drivers/iio/light/si1133.c line 90
Quoted 9 lines
drivers/iio/light/si1133.c +0
-2#define SI1133_CMD_PARAM_SET 0x80#define SI1133_CMD_PARAM_QUERY 0x40#define SI1133_CMD_PARAM_MASK 0x3F@@ -86,13 +83,9 @@#define SI1133_CMD_MINSLEEP_US_LOW 5000#define SI1133_CMD_MINSLEEP_US_HIGH 7500#define SI1133_CMD_TIMEOUT_MS 25 -#define SI1133_CMD_LUX_TIMEOUT_MS 5000 -#define SI1133_CMD_TIMEOUT_US (SI1133_CMD_TIMEOUT_MS * 1000)
Dropping these and the one below is good as those aren't register or command defines. This last one is entirely pointless as better to have the rescale inline and only one definition up here.
Quoted 9 lines
drivers/iio/light/si1133.c +0
-2#define SI1133_REG_HOSTOUT (x) ((x) + 0x13) -#define SI1133_MEASUREMENT_FREQUENCY 1250 -#define SI1133_X_ORDER_MASK 0x0070#define SI1133_Y_ORDER_MASK 0x0007#define si1133_get_x_order (m) (((m) & SI1133_X_ORDER_MASK) >> 4)
Re: [PATCH v2 2/5] iio: light: si1133: remove unused macros
Quoted 2 lines · Jonathan Cameron
Don't update macros then remove them - that's just increasing noise in the update patch. Reorder this patch to be before patch 1.
Yeah, my fault, I forgot to do a rebase.
Quoted 22 lines
drivers/iio/light/si1133.c line 26
Quoted 5 lines
#define SI1133_REG_PART_ID 0x00 #define SI1133_REG_REV_ID 0x01 #define SI1133_REG_MFR_ID 0x02drivers/iio/light/si1133.c +0
-2-#define SI1133_REG_INFO0 0x03 -#define SI1133_REG_INFO1 0x04This one isn't in the category of things Andy was referring to.
It's a hardware register address definition. We don't have hard rules on whether a complete set of these should be defined, or just the ones used but it is definitely acceptable to just define them all up front as then easy to check them against a datasheet.drivers/iio/light/si1133.c line 43
Quoted 8 lines
#define
SI1133_PART_ID0x33drivers/iio/light/si1133.c +0
-1@@ -40,7 +38,6 @@#define SI1133_CMD_RESET_CTR 0x00#define SI1133_CMD_RESET_SW 0x01#define SI1133_CMD_FORCE 0x11 -#define SI1133_CMD_START_AUTONOMOUS 0x13Similar for this.
Quoted 13 lines
drivers/iio/light/si1133.c line 90
Quoted 9 lines
drivers/iio/light/si1133.c +0
-2#define SI1133_CMD_PARAM_SET 0x80#define SI1133_CMD_PARAM_QUERY 0x40#define SI1133_CMD_PARAM_MASK 0x3F@@ -86,13 +83,9 @@#define SI1133_CMD_MINSLEEP_US_LOW 5000#define SI1133_CMD_MINSLEEP_US_HIGH 7500#define SI1133_CMD_TIMEOUT_MS 25 -#define SI1133_CMD_LUX_TIMEOUT_MS 5000 -#define SI1133_CMD_TIMEOUT_US (SI1133_CMD_TIMEOUT_MS * 1000)Dropping these and the one below is good as those aren't register or command defines. This last one is entirely pointless as better to have the rescale inline and only one definition up here.
Sure, I'll send a v2 tomorrow.
Signature · 3 lines
Kind regards CJD
PATCH v2 3/5 iio: light: si1133: add missing include headers
In reply to Joshua CroftsAdd missing include headers to prevent compilation relying on transient dependencies (array_size.h, bitops.h, completion.h, dev_printk.h, err.h, irqreturn.h, jiffies.h, math.h, mutex.h, types.h).
--- drivers/iio/light/si1133.c | 10 ++++++++++ 1 file changed, 10 insertions(+)
drivers/iio/light/si1133.c +10 -0
@@ -6,11 +6,21 @@* Copyright 2018 Maxime Roussin-Belanger <[email protected]> */ +#include <linux/array_size.h> +#include <linux/bitops.h> +#include <linux/completion.h> #include <linux/delay.h> +#include <linux/dev_printk.h> +#include <linux/err.h> #include <linux/i2c.h> #include <linux/interrupt.h> +#include <linux/irqreturn.h> +#include <linux/jiffies.h> +#include <linux/math.h> #include <linux/module.h> +#include <linux/mutex.h> #include <linux/regmap.h> +#include <linux/types.h> #include <linux/iio/iio.h> #include <linux/iio/sysfs.h>
Signature · 1 line
2.47.3
PATCH v2 4/5 iio: light: si1133: group generic <linux/*> headers
In reply to Joshua CroftsGroup generic <linux/*> include headers to improve code style.
No functional change.
--- drivers/iio/light/si1133.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-)
drivers/iio/light/si1133.c +2 -4
@@ -21,14 +21,12 @@#include <linux/mutex.h> #include <linux/regmap.h> #include <linux/types.h> +#include <linux/unaligned.h> +#include <linux/util_macros.h> #include <linux/iio/iio.h> #include <linux/iio/sysfs.h> -#include <linux/util_macros.h> - -#include <linux/unaligned.h> -#define SI1133_REG_PART_ID 0x00#define SI1133_REG_REV_ID 0x01#define SI1133_REG_MFR_ID 0x02
Signature · 1 line
2.47.3
Jonathan Cameron
maintainerRe: [PATCH v2 4/5] iio: light: si1133: group generic <linux/*> headers
Quoted 3 lines · On Tue, 28 Apr 2026 15:18:36 +0200 Joshua Crofts via B4 Relay…
From: Joshua Crofts <[email protected]>
Group generic <linux/*> include headers to improve code style.
Not particularly important but you could mention why these two are where they are. For the unaligned one that was a mass replacement of asm/unaligned with linux/unaligned that didn't reflect local ordering. No idea for util_macros.h!
Quoted 30 lines
No functional change.
Signed-off-by Joshua Crofts <[email protected]>--- drivers/iio/light/si1133.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-)drivers/iio/light/si1133.c +2
-4@@ -21,14 +21,12 @@#include <linux/mutex.h> #include <linux/regmap.h> #include <linux/types.h> +#include <linux/unaligned.h> +#include <linux/util_macros.h> #include <linux/iio/iio.h> #include <linux/iio/sysfs.h> -#include <linux/util_macros.h> - -#include <linux/unaligned.h> -#define SI1133_REG_PART_ID 0x00#define SI1133_REG_REV_ID 0x01#define SI1133_REG_MFR_ID 0x02
PATCH v2 5/5 iio: light: si1133: use guard(mutex)() macro
In reply to Joshua CroftsRemove mutex_lock()/mutex_unlock() and goto instances and add guard(mutex)() macro to modernize driver and improve mutex handling.
--- drivers/iio/light/si1133.c | 18 +++++++----------- 1 file changed, 7 insertions(+), 11 deletions(-)
drivers/iio/light/si1133.c +7 -11
@@ -8,6 +8,7 @@#include <linux/array_size.h> #include <linux/bitops.h> +#include <linux/cleanup.h> #include <linux/completion.h> #include <linux/delay.h> #include <linux/dev_printk.h>
@@ -392,7 +393,7 @@ static int si1133_command(struct si1133_data *data, u8 cmd)int err;int expected_seq; -mutex_lock (&data->mutex); +guard (mutex)(&data->mutex); expected_seq = (data->rsp_seq +1 ) &SI1133_MAX_CMD_CTR ;
@@ -403,19 +404,17 @@ static int si1133_command(struct si1133_data *data, u8 cmd)if (err) {dev_warn (dev,"Failed to write command 0x%02x, ret=%d\n" , cmd, err); - goto out; +return err; }if (cmd ==SI1133_CMD_FORCE ) {/* wait for irq */ if (!wait_for_completion_timeout (&data->completion, -msecs_to_jiffies (SI1133_COMPLETION_TIMEOUT_MS ))) { - err = -ETIMEDOUT ; - goto out; - } +msecs_to_jiffies (SI1133_COMPLETION_TIMEOUT_MS ))) +return -ETIMEDOUT ; err =regmap_read (data->regmap,SI1133_REG_RESPONSE0 , &resp);if (err) - goto out; +return err; }else { err = regmap_read_poll_timeout(data->regmap, SI1133_REG_RESPONSE0, resp,
@@ -428,7 +427,7 @@ static int si1133_command(struct si1133_data *data, u8 cmd)dev_warn (dev,"Failed to read command 0x%02x, ret=%d\n" , cmd, err); - goto out; +return err; } }
@@ -439,9 +438,6 @@ static int si1133_command(struct si1133_data *data, u8 cmd) data->rsp_seq = expected_seq; } -out: -mutex_unlock (&data->mutex); -return err; }
Signature · 1 line
2.47.3
Jonathan Cameron
maintainerRe: [PATCH v2 5/5] iio: light: si1133: use guard(mutex)() macro
Quoted 6 lines · On Tue, 28 Apr 2026 15:18:37 +0200 Joshua Crofts via B4 Relay…
From: Joshua Crofts <[email protected]>
Remove mutex_lock()/mutex_unlock() and goto instances and add guard(mutex)() macro to modernize driver and improve mutex handling.
Signed-off-by Joshua Crofts <[email protected]>
One comment that is more about the existing code than the changes here, but as you are touching it nice to clean that bit up!
Nice series,
Jonathan
Quoted 32 lines
--- drivers/iio/light/si1133.c | 18 +++++++----------- 1 file changed, 7 insertions(+), 11 deletions(-)drivers/iio/light/si1133.c +3
-2@@ -8,6 +8,7 @@#include <linux/array_size.h> #include <linux/bitops.h> +#include <linux/cleanup.h> #include <linux/completion.h> #include <linux/delay.h> #include <linux/dev_printk.h> @@ -392,7 +393,7 @@ static int si1133_command(struct si1133_data *data, u8 cmd)int err;int expected_seq; -mutex_lock (&data->mutex); +guard (mutex)(&data->mutex); expected_seq = (data->rsp_seq +1 ) &SI1133_MAX_CMD_CTR ;@@ -403,19 +404,17 @@ static int si1133_command(struct si1133_data *data, u8 cmd)if (err) {dev_warn (dev,"Failed to write command 0x%02x, ret=%d\n" , cmd, err); - goto out; +return err; }
drivers/iio/light/si1133.c static int si1133_command(struct si1133_data *data, u8 cmd) · line 414
Quoted 10 lines
drivers/iio/light/si1133.c +2
-4if (cmd ==SI1133_CMD_FORCE ) {/* wait for irq */ if (!wait_for_completion_timeout (&data->completion, -msecs_to_jiffies (SI1133_COMPLETION_TIMEOUT_MS ))) { - err = -ETIMEDOUT ; - goto out; - } +msecs_to_jiffies (SI1133_COMPLETION_TIMEOUT_MS ))) +return -ETIMEDOUT ;
This is some unfortunate indentation from readability point of view.
I'd reindent and not worry about the long line - it's very long but still under 100 so tools won't complain and I think it is worth the pain here!
drivers/iio/light/si1133.c static int si1133_command(struct si1133_data *data, u8 cmd) · line 412
Quoted 1 line
drivers/iio/light/si1133.c +0
-0if (!wait_for_completion_timeout(&data->completion,
msecs_to_jiffies(SI1133_COMPLETION_TIMEOUT_MS))) return -ETIMEDOUT;
or perhaps better yet, use a local variable for that timeout.
Quoted 27 lines
drivers/iio/light/si1133.c +1
-1err =regmap_read (data->regmap,SI1133_REG_RESPONSE0 , &resp);if (err) - goto out; +return err; }else { err = regmap_read_poll_timeout(data->regmap,drivers/iio/light/si1133.c +1
-4SI1133_REG_RESPONSE0, resp,@@ -428,7 +427,7 @@ static int si1133_command(struct si1133_data *data, u8 cmd)dev_warn (dev,"Failed to read command 0x%02x, ret=%d\n" , cmd, err); - goto out; +return err; } }@@ -439,9 +438,6 @@ static int si1133_command(struct si1133_data *data, u8 cmd) data->rsp_seq = expected_seq; } -out: -mutex_unlock (&data->mutex); -return err; }
Re: [PATCH v2 5/5] iio: light: si1133: use guard(mutex)() macro
Quoted 13 lines · Jonathan Cameron
drivers/iio/light/si1133.c static int si1133_command(struct si1133_data *data, u8 cmd) · line 414
Quoted 9 lines
if (cmd == SI1133_CMD_FORCE) { /* wait for irq */ if (!wait_for_completion_timeout(&data->completion,drivers/iio/light/si1133.c +2
-4- msecs_to_jiffies(SI1133_COMPLETION_TIMEOUT_MS))) { - err = -ETIMEDOUT ; - goto out; - } + msecs_to_jiffies(SI1133_COMPLETION_TIMEOUT_MS))) +return -ETIMEDOUT ;This is some unfortunate indentation from readability point of view.
I'd reindent and not worry about the long line - it's very long but still under 100 so tools won't complain and I think it is worth the pain here!
Quoted 5 lines
Quoted 1 line
if (!wait_for_completion_timeout(&data->completion,msecs_to_jiffies(SI1133_COMPLETION_TIMEOUT_MS))) return -ETIMEDOUT;or perhaps better yet, use a local variable for that timeout.
Okay, I'll probably do a separate patch for that though.
Signature · 3 lines
Kind regards CJD