Skip to contentSkip to conversation

iio: core: Replace BUG() with WARN_ON_ONCE() and error return

PATCH In review Rishab Madhugiri 3 participants 4 messages Aug 26 – Aug 30 IIO

lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 iio: core: Replace BUG() with WARN_ON_ONCE() and error return

As documented in Documentation/process/deprecated.rst, the use of BUG() and BUG_ON() should be avoided as they could cause a complete system crash, preventing further debugging.

In iio_get_time_ns() and current_timestamp_clock_show(), the clock type is already validated against supported clock types prior to these calls, making the default switch branches expected to be unreachable. Replace these with WARN_ON_ONCE(1) followed by an error return to avoid a crash and to log the highly unexpected condition.

---
v2:
- Use WARN_ON_ONCE(1) instead of WARN_ONCE() with string message as suggested by Jonathan Cameron.
- Reference Documentation/process/deprecated.rst and clarify unreachability of default branches and reasoning for using WARN family over BUG family in commit message as per Andy Shevchenko's feedback.

 drivers/iio/industrialio-core.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/iio/industrialio-core.c +4 -2
index 767a7794624a..819a864d1136 100644
@@ -341,7 +341,8 @@ s64 iio_get_time_ns(const struct iio_dev *indio_dev)
 	case CLOCK_TAI:
 		return ktime_get_clocktai_ns();
 	default:
-		BUG();
+		WARN_ON_ONCE(1);
+		return 0;
 	}
 }
 EXPORT_SYMBOL(iio_get_time_ns);
@@ -1518,7 +1519,8 @@ static ssize_t current_timestamp_clock_show(struct device *dev,
 	case CLOCK_TAI:
 		break;
 	default:
-		BUG();
+		WARN_ON_ONCE(1);
+		return -EINVAL;
 	}
 
 	return sysfs_emit(buf, "%s\n", clock_names[clk]);
Signature · 1 line
2.43.0
#1 lore raw
Quoted 3 lines · Rishab Madhugiri

As documented in Documentation/process/deprecated.rst, the use of BUG() and BUG_ON() should be avoided as they could cause a complete system crash, preventing further debugging.

Right, but deprecation doesn't mean the _old_ existing code must be immediately changed. I leave it to Jonathan, but I see no value in this change except the unneeded churn. If we ever want to drop BUG(), it should be done tree wide. Hence no tag given.

Quoted 6 lines

In iio_get_time_ns() and current_timestamp_clock_show(), the clock type is already validated against supported clock types prior to these calls, making the default switch branches expected to be unreachable. Replace these with WARN_ON_ONCE(1) followed by an error return to avoid a crash and to log the highly unexpected condition.

There is like 10+ character space on each line, the commit message shouldn't be that condensed.

Signature · 2 lines
With Best Regards,
Andy Shevchenko
author
#2 lore raw

Thanks for the feedback. I totally understand the perspective on code churn versus making tree wide cleanups, and I will let Jonathan make the final call on whether this is suitable. I've also noted the wrapping width feedback for commit messages in the future.

Best,
Rishab

Quoted 26 lines · Andy Shevchenko
Quoted 3 lines · Rishab Madhugiri

As documented in Documentation/process/deprecated.rst, the use of BUG() and BUG_ON() should be avoided as they could cause a complete system crash, preventing further debugging.

Right, but deprecation doesn't mean the _old_ existing code must be immediately changed. I leave it to Jonathan, but I see no value in this change except the unneeded churn. If we ever want to drop BUG(), it should be done tree wide. Hence no tag given.

Quoted 6 lines

In iio_get_time_ns() and current_timestamp_clock_show(), the clock type is already validated against supported clock types prior to these calls, making the default switch branches expected to be unreachable. Replace these with WARN_ON_ONCE(1) followed by an error return to avoid a crash and to log the highly unexpected condition.

There is like 10+ character space on each line, the commit message shouldn't be that condensed.

--
With Best Regards,
Andy Shevchenko

maintainer
#3 lore raw
Quoted 5 lines · On Thu, 27 Aug 2026 15:24:52 -0700 Rishab M <[email protected]> wrote:

Thanks for the feedback. I totally understand the perspective on code churn versus making tree wide cleanups, and I will let Jonathan make the final call on whether this is suitable. I've also noted the wrapping width feedback for commit messages in the future.

Please comment inline.

Anyhow, I have applied this but mostly because it is easier for cases like this to do so than to reply to a series of folk proposing the change.

Slightly rubbish reason, but meh, too many patches floating around and this is a harmless way to potentially cut that down a little in the longer term.

Thanks,

Jonathan

Quoted 31 lines

Best,
Rishab

Quoted 26 lines · Andy Shevchenko
Quoted 3 lines · Rishab Madhugiri

As documented in Documentation/process/deprecated.rst, the use of BUG() and BUG_ON() should be avoided as they could cause a complete system crash, preventing further debugging.

Right, but deprecation doesn't mean the _old_ existing code must be immediately changed. I leave it to Jonathan, but I see no value in this change except the unneeded churn. If we ever want to drop BUG(), it should be done tree wide. Hence no tag given.

Quoted 6 lines

In iio_get_time_ns() and current_timestamp_clock_show(), the clock type is already validated against supported clock types prior to these calls, making the default switch branches expected to be unreachable. Replace these with WARN_ON_ONCE(1) followed by an error return to avoid a crash and to log the highly unexpected condition.

There is like 10+ character space on each line, the commit message shouldn't be that condensed.

--
With Best Regards,
Andy Shevchenko