Skip to contentSkip to conversation

compiler_types: Warn about unused static inline functions on second

PATCH Applied Andy Shevchenko 3 participants 7 messages Nov 6, 2025 – Nov 7, 2025

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 compiler_types: Warn about unused static inline functions on second

From: Peter Zijlstra <[email protected]>

Per Nathan, clang catches unused "static inline" functions in C files since commit 6863f5643dd7 ("kbuild: allow Clang to find unused static inline functions for W=1 build").

Quoted 8 lines · Linus

So I entirely ignore W=1 issues, because I think so many of the extra warnings are bogus.

But if this one in particular is causing more problems than most - some teams do seem to use W=1 as part of their test builds - it's fine to send me a patch that just moves bad warnings to W=2.

And if anybody uses W=2 for their test builds, that's THEIR problem..

Here is the change to bump the warning from W=1 to W=2.

Fixes 6863f5643dd7 ("kbuild: allow Clang to find unused static inline functions for W=1 build")
---
 include/linux/compiler_types.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
include/linux/compiler_types.h +1 -1
index 59288a2c1ad2..575a03643fa1 100644
@@ -253,7 +253,7 @@ struct ftrace_likely_data {
  * for W=1 build. This will allow clang to find unused functions. Remove the
  * __inline_maybe_unused entirely after fixing most of -Wunused-function warnings.
  */
-#ifdef KBUILD_EXTRA_WARN1
+#ifdef KBUILD_EXTRA_WARN2
 #define __inline_maybe_unused
 #else
 #define __inline_maybe_unused __maybe_unused
Signature · 1 line
2.50.1
#1 lore raw
Quoted 15 lines · Andy Shevchenko

From: Peter Zijlstra <[email protected]>

Per Nathan, clang catches unused "static inline" functions in C files since commit 6863f5643dd7 ("kbuild: allow Clang to find unused static inline functions for W=1 build").

Quoted 7 lines · Linus

So I entirely ignore W=1 issues, because I think so many of the extra warnings are bogus.

But if this one in particular is causing more problems than most - some teams do seem to use W=1 as part of their test builds - it's fine to send me a patch that just moves bad warnings to W=2.

Quoted 7 lines
Quoted 1 line

And if anybody uses W=2 for their test builds, that's THEIR problem..

Here is the change to bump the warning from W=1 to W=2.

Fixes 6863f5643dd7 ("kbuild: allow Clang to find unused static inline functions for W=1 build")

I can take this as a fix via kbuild or Linus can apply it directly but...

include/linux/compiler_types.h struct ftrace_likely_data { · line 255

Quoted 12 lines
---
 include/linux/compiler_types.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
include/linux/compiler_types.h +0 -0
index 59288a2c1ad2..575a03643fa1 100644
@@ -253,7 +253,7 @@ struct ftrace_likely_data {
  * for W=1 build. This will allow clang to find unused functions. Remove the
  * __inline_maybe_unused entirely after fixing most of -Wunused-function warnings.
  */

The comment should be updated to W=2 instead of W=1 and we should probably drop the sentence about removing __inline_maybe_unused entirely since people such as Peter will never want this behavior by default. I do not mind doing it myself if I take it.

Quoted 8 lines
include/linux/compiler_types.h +1 -1
-#ifdef KBUILD_EXTRA_WARN1
+#ifdef KBUILD_EXTRA_WARN2
 #define __inline_maybe_unused
 #else
 #define __inline_maybe_unused __maybe_unused
Signature · 1 line
2.50.1
#2 lore raw
Quoted 1 line · Nathan Chancellor

On Thu, Nov 06, 2025 at 11:50:00AM +0100, Andy Shevchenko wrote:

...

Quoted 4 lines
Quoted 2 lines

* for W=1 build. This will allow clang to find unused functions. Remove the
* __inline_maybe_unused entirely after fixing most of -Wunused-function warnings.

The comment should be updated to W=2 instead of W=1

Granted.

Quoted 4 lines

and we should
probably drop the sentence about removing __inline_maybe_unused entirely since people such as Peter will never want this behavior by default. I do not mind doing it myself if I take it.

But future is uncertain, it might be that GCC also gains this and it won't confuse anyway as it might become a truth (no more such warnings in the code) at some point.

Signature · 2 lines
With Best Regards,
Andy Shevchenko
#3 lore raw
Quoted 9 lines · Andy Shevchenko
Quoted 4 lines · Nathan Chancellor

and we should
probably drop the sentence about removing __inline_maybe_unused entirely since people such as Peter will never want this behavior by default. I do not mind doing it myself if I take it.

But future is uncertain, it might be that GCC also gains this and it won't confuse anyway as it might become a truth (no more such warnings in the code) at some point.

The reality of the situation is that moving this warning to W=2 is basically the same as just turning it off entirely since building with W=2 is not a common endeavor for the majority of folks actually writing kernel code, so the number of warnings will just continue to grow. At that point, there is very little reason to believe that we would be able to go from W=2 to enabled by default at some point in the future since people already do not like it enabled at W=1 where it is not as impactful as enabled by default. As a result, I feel like the comment genuinely serves no purpose. If GCC were to change its behavior to match clang, I feel like kernel folks would still want the current GCC behavior.

Cheers,
Nathan

#4 lore raw
Quoted 11 lines · Nathan Chancellor
Quoted 9 lines · Andy Shevchenko
Quoted 4 lines · Nathan Chancellor

and we should
probably drop the sentence about removing __inline_maybe_unused entirely since people such as Peter will never want this behavior by default. I do not mind doing it myself if I take it.

But future is uncertain, it might be that GCC also gains this and it won't confuse anyway as it might become a truth (no more such warnings in the code) at some point.

Quoted 11 lines

The reality of the situation is that moving this warning to W=2 is basically the same as just turning it off entirely since building with W=2 is not a common endeavor for the majority of folks actually writing kernel code, so the number of warnings will just continue to grow. At that point, there is very little reason to believe that we would be able to go from W=2 to enabled by default at some point in the future since people already do not like it enabled at W=1 where it is not as impactful as enabled by default. As a result, I feel like the comment genuinely serves no purpose. If GCC were to change its behavior to match clang, I feel like kernel folks would still want the current GCC behavior.

Okay, I won't object. But so far the patch taken as is, please fold the changes you mention in it (will require rebase, but personally I don't care, all depends on your workflow, followup is also fine to me).

And thanks for taking care of this!

Signature · 2 lines
With Best Regards,
Andy Shevchenko
#5 lore raw
In reply to Nathan Chancellor
Quoted 2 lines · Nathan Chancellor

I can take this as a fix via kbuild or Linus can apply it directly

Yes, please just take it through the regular Kbuild tree (with whatever modifications this thread then results in).

Thanks,
Linus

#6 lore raw
In reply to Andy Shevchenko
Quoted 9 lines · Andy Shevchenko

Per Nathan, clang catches unused "static inline" functions in C files since commit 6863f5643dd7 ("kbuild: allow Clang to find unused static inline functions for W=1 build").

Quoted 3 lines · Linus

So I entirely ignore W=1 issues, because I think so many of the extra warnings are bogus.

Quoted 7 lines
Quoted 5 lines

But if this one in particular is causing more problems than most - some teams do seem to use W=1 as part of their test builds - it's fine to send me a patch that just moves bad warnings to W=2.

And if anybody uses W=2 for their test builds, that's THEIR problem..

[...]

Applied, thanks!

[1/1] compiler_types: Warn about unused static inline functions on second

      https://git.kernel.org/kbuild/c/9818af18db4bf

Best regards,

Signature · 1 line
Nathan Chancellor <[email protected]>