Skip to contentSkip to conversation

leds: turris-omnia: Fix unused variable

PATCH In review Marek Behún 2 participants 4 messages Sep 21, 2023 – Sep 25, 2023

Subsystems (2)
lore
Share & tools
lore

Marek Behún

author maintainer
#0 lore raw

PATCH 1/1 leds: turris-omnia: Fix unused variable

The variable ret is not used in this function.

Fixes 28350bc0ac77 ("leds: turris-omnia: Do not use SMBUS calls")
---
 drivers/leds/leds-turris-omnia.c | 1 -
 1 file changed, 1 deletion(-)
drivers/leds/leds-turris-omnia.c +0 -1
index e1a4629479c5..f27241896970 100644
@@ -60,7 +60,6 @@ struct omnia_leds {
 static int omnia_cmd_write_u8(const struct i2c_client *client, u8 cmd, u8 val)
 {
 	u8 buf[2] = { cmd, val };
-	int ret;
 
 	return i2c_master_send(client, buf, sizeof(buf));
 }
Signature · 1 line
2.41.0
maintainer
#1 lore raw
Quoted 8 lines · Marek Behún

The variable ret is not used in this function.

Fixes 28350bc0ac77 ("leds: turris-omnia: Do not use SMBUS calls")
---
 drivers/leds/leds-turris-omnia.c | 1 -
 1 file changed, 1 deletion(-)

I already fixed and squashed this.

How was this missed when you tested the set?

Signature · 1 line
Lee Jones [李琼斯]
author maintainer
#2 lore raw
Quoted 11 lines · On Fri, 22 Sep 2023 07:59:19 +0100 Lee Jones <[email protected]> wrote:
Quoted 8 lines · Marek Behún

The variable ret is not used in this function.

Fixes 28350bc0ac77 ("leds: turris-omnia: Do not use SMBUS calls")
---
 drivers/leds/leds-turris-omnia.c | 1 -
 1 file changed, 1 deletion(-)  
Quoted 3 lines

I already fixed and squashed this.

How was this missed when you tested the set?

I am not sure, but it is possible that I've refactored that function from my original (return 0 on success) to your proposed (return number of written bytes on success) and did not notice the warning. I am sure I tested that the result works. Maybe I switched to another terminal where I was testing it too fast and did not notice the warning.

Sorry about this.

Anyway, I have a question. Several days ago I also sent for review a new driver for other feautres the Turris Omnia MCU provides (GPIO, watchdog, wakeup+poweroff).
There, I also refactored the _write and _read functions as you suggested (to return the number of bytes written/read).
On review, Andy Shevchenko requested [1] to refactor it to my original (return 0 on success). I mentioned to him [2] your request, to which he replied [3]:
This is strange. For example, regmap APIs never returns amount of data written or read. I think it's solely depends on the API. It might be useful for i²c APIs, in case you can do something about it. but if you have wrappers on top of that already (meaning not using directly the i2c_*() calls, I dunno the positive return is anyhow useful.
Since I agree with him, taking this into account, would you accept a patch that returns those function to how I originally wrote them (return 0 on success)?

Thanks.

Marek

[1]
https://lore.kernel.org/linux-gpio/[email protected]/
[2]
https://lore.kernel.org/linux-gpio/[email protected]/
[3]
https://lore.kernel.org/linux-gpio/[email protected]/

maintainer
#3 lore raw
Quoted 26 lines · Marek Behún
Quoted 11 lines · On Fri, 22 Sep 2023 07:59:19 +0100 Lee Jones <[email protected]> wrote:
Quoted 8 lines · Marek Behún

The variable ret is not used in this function.

Fixes 28350bc0ac77 ("leds: turris-omnia: Do not use SMBUS calls")
---
 drivers/leds/leds-turris-omnia.c | 1 -
 1 file changed, 1 deletion(-)  
Quoted 3 lines

I already fixed and squashed this.

How was this missed when you tested the set?

I am not sure, but it is possible that I've refactored that function from my original (return 0 on success) to your proposed (return number of written bytes on success) and did not notice the warning. I am sure I tested that the result works. Maybe I switched to another terminal where I was testing it too fast and did not notice the warning.

Sorry about this.

Quoted 16 lines

Anyway, I have a question. Several days ago I also sent for review a new driver for other feautres the Turris Omnia MCU provides (GPIO, watchdog, wakeup+poweroff).
There, I also refactored the _write and _read functions as you suggested (to return the number of bytes written/read).
On review, Andy Shevchenko requested [1] to refactor it to my original (return 0 on success). I mentioned to him [2] your request, to which he replied [3]:
This is strange. For example, regmap APIs never returns amount of data written or read. I think it's solely depends on the API. It might be useful for i²c APIs, in case you can do something about it. but if you have wrappers on top of that already (meaning not using directly the i2c_*() calls, I dunno the positive return is anyhow useful.
Since I agree with him, taking this into account, would you accept a patch that returns those function to how I originally wrote them (return 0 on success)?

As I said before, I'm not going to force you into anything.

Fire away.

Signature · 1 line
Lee Jones [李琼斯]