[PATCH] seqlock: Allow UBSAN_ALIGNMENT to fail optimizing

[PATCH] seqlock: Allow UBSAN_ALIGNMENT to fail optimizing

[Heiko Carstens]

With gcc-15 and gcc-16 with UBSAN_ALIGNMENT enabled the compiler fails to
inline and optimize __scoped_seqlock_bug() away on s390:

s390x-16.1.0-ld: kernel/sched/build_policy.o: in function `__scoped_seqlock_next':
/.../seqlock.h:1286:(.text+0x22030): undefined reference to `__scoped_seqlock_bug'

Fix this by adding UBSAN_ALIGNMENT to the list of config options where a
not inlined empty __scoped_seqlock_bug() is allowed.

Reported-by: Arnd Bergmann <arnd@arndb.de>
Closes: https://lore.kernel.org/r/20260515092057.810542-1-arnd@kernel.org/
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
---
 include/linux/seqlock.h | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/include/linux/seqlock.h b/include/linux/seqlock.h
index 5a40252b8334..f865491c4f2c 100644
--- a/include/linux/seqlock.h
+++ b/include/linux/seqlock.h
@@ -1259,14 +1259,15 @@ static __always_inline void __scoped_seqlock_cleanup(struct ss_tmp *sst)
 
 extern void __scoped_seqlock_invalid_target(void);
 
-#if (defined(CONFIG_CC_IS_GCC) && CONFIG_GCC_VERSION < 90000) || defined(CONFIG_KASAN)
+#if (defined(CONFIG_CC_IS_GCC) && CONFIG_GCC_VERSION < 90000) || \
+	defined(CONFIG_KASAN) || defined(CONFIG_UBSAN_ALIGNMENT)
 /*
  * For some reason some GCC-8 architectures (nios2, alpha) have trouble
  * determining that the ss_done state is impossible in __scoped_seqlock_next()
  * below.
  *
- * Similarly KASAN is known to confuse compilers enough to break this. But we
- * don't care about code quality for KASAN builds anyway.
+ * Similarly KASAN and UBSAN_ALIGNMENT are known to confuse compilers enough
+ * to break this. But we don't care about code quality for such builds anyway.
  */
 static inline void __scoped_seqlock_bug(void) { }
 #else
-- 
2.51.0

Re: [PATCH] seqlock: Allow UBSAN_ALIGNMENT to fail optimizing

[Arnd Bergmann]

On Tue, May 19, 2026, at 13:03, Heiko Carstens wrote:
> With gcc-15 and gcc-16 with UBSAN_ALIGNMENT enabled the compiler fails to
> inline and optimize __scoped_seqlock_bug() away on s390:

technically

s/on s390//

since I have been able to reproduce the problem across architectures
now.

> s390x-16.1.0-ld: kernel/sched/build_policy.o: in function 
> `__scoped_seqlock_next':
> /.../seqlock.h:1286:(.text+0x22030): undefined reference to 
> `__scoped_seqlock_bug'
>
> Fix this by adding UBSAN_ALIGNMENT to the list of config options where a
> not inlined empty __scoped_seqlock_bug() is allowed.
>
> Reported-by: Arnd Bergmann <arnd@arndb.de>
> Closes: https://lore.kernel.org/r/20260515092057.810542-1-arnd@kernel.org/
> Signed-off-by: Heiko Carstens <hca@linux.ibm.com>

I was about to send the same patch.

Tested-by: Arnd Bergmann <arnd@arndb.de>

      Arnd

[tip: locking/core] seqlock: Allow UBSAN_ALIGNMENT to fail optimizing

[tip-bot2 for Heiko Carstens]

The following commit has been merged into the locking/core branch of tip:

Commit-ID:     43a037d4fa6d7e7f9437e069f857a5e06eb2dff7
Gitweb:        https://git.kernel.org/tip/43a037d4fa6d7e7f9437e069f857a5e06eb2dff7
Author:        Heiko Carstens <hca@linux.ibm.com>
AuthorDate:    Tue, 19 May 2026 13:03:15 +02:00
Committer:     Peter Zijlstra <peterz@infradead.org>
CommitterDate: Tue, 19 May 2026 13:49:01 +02:00

seqlock: Allow UBSAN_ALIGNMENT to fail optimizing

With gcc-15 and gcc-16 with UBSAN_ALIGNMENT enabled the compiler fails to
inline and optimize __scoped_seqlock_bug() away on s390:

s390x-16.1.0-ld: kernel/sched/build_policy.o: in function `__scoped_seqlock_next':
/.../seqlock.h:1286:(.text+0x22030): undefined reference to `__scoped_seqlock_bug'

Fix this by adding UBSAN_ALIGNMENT to the list of config options where a
not inlined empty __scoped_seqlock_bug() is allowed.

Closes: https://lore.kernel.org/r/20260515092057.810542-1-arnd@kernel.org/
Reported-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://patch.msgid.link/20260519110315.1385307-1-hca@linux.ibm.com
---
 include/linux/seqlock.h | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/include/linux/seqlock.h b/include/linux/seqlock.h
index 5a40252..f865491 100644
--- a/include/linux/seqlock.h
+++ b/include/linux/seqlock.h
@@ -1259,14 +1259,15 @@ static __always_inline void __scoped_seqlock_cleanup(struct ss_tmp *sst)
 
 extern void __scoped_seqlock_invalid_target(void);
 
-#if (defined(CONFIG_CC_IS_GCC) && CONFIG_GCC_VERSION < 90000) || defined(CONFIG_KASAN)
+#if (defined(CONFIG_CC_IS_GCC) && CONFIG_GCC_VERSION < 90000) || \
+	defined(CONFIG_KASAN) || defined(CONFIG_UBSAN_ALIGNMENT)
 /*
  * For some reason some GCC-8 architectures (nios2, alpha) have trouble
  * determining that the ss_done state is impossible in __scoped_seqlock_next()
  * below.
  *
- * Similarly KASAN is known to confuse compilers enough to break this. But we
- * don't care about code quality for KASAN builds anyway.
+ * Similarly KASAN and UBSAN_ALIGNMENT are known to confuse compilers enough
+ * to break this. But we don't care about code quality for such builds anyway.
  */
 static inline void __scoped_seqlock_bug(void) { }
 #else

[tip: locking/core] seqlock: Allow UBSAN_ALIGNMENT to fail optimizing

[tip-bot2 for Heiko Carstens]

The following commit has been merged into the locking/core branch of tip:

Commit-ID:     88331c4ec23a28c1006ec532fa64763d4c695e90
Gitweb:        https://git.kernel.org/tip/88331c4ec23a28c1006ec532fa64763d4c695e90
Author:        Heiko Carstens <hca@linux.ibm.com>
AuthorDate:    Tue, 19 May 2026 13:03:15 +02:00
Committer:     Peter Zijlstra <peterz@infradead.org>
CommitterDate: Thu, 28 May 2026 12:23:22 +02:00

seqlock: Allow UBSAN_ALIGNMENT to fail optimizing

With gcc-15 and gcc-16 with UBSAN_ALIGNMENT enabled the compiler fails to
inline and optimize __scoped_seqlock_bug() away on s390:

s390x-16.1.0-ld: kernel/sched/build_policy.o: in function `__scoped_seqlock_next':
/.../seqlock.h:1286:(.text+0x22030): undefined reference to `__scoped_seqlock_bug'

Fix this by adding UBSAN_ALIGNMENT to the list of config options where a
not inlined empty __scoped_seqlock_bug() is allowed.

Closes: https://lore.kernel.org/r/20260515092057.810542-1-arnd@kernel.org/
Reported-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://patch.msgid.link/20260519110315.1385307-1-hca@linux.ibm.com
---
 include/linux/seqlock.h | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/include/linux/seqlock.h b/include/linux/seqlock.h
index 5a40252..f865491 100644
--- a/include/linux/seqlock.h
+++ b/include/linux/seqlock.h
@@ -1259,14 +1259,15 @@ static __always_inline void __scoped_seqlock_cleanup(struct ss_tmp *sst)
 
 extern void __scoped_seqlock_invalid_target(void);
 
-#if (defined(CONFIG_CC_IS_GCC) && CONFIG_GCC_VERSION < 90000) || defined(CONFIG_KASAN)
+#if (defined(CONFIG_CC_IS_GCC) && CONFIG_GCC_VERSION < 90000) || \
+	defined(CONFIG_KASAN) || defined(CONFIG_UBSAN_ALIGNMENT)
 /*
  * For some reason some GCC-8 architectures (nios2, alpha) have trouble
  * determining that the ss_done state is impossible in __scoped_seqlock_next()
  * below.
  *
- * Similarly KASAN is known to confuse compilers enough to break this. But we
- * don't care about code quality for KASAN builds anyway.
+ * Similarly KASAN and UBSAN_ALIGNMENT are known to confuse compilers enough
+ * to break this. But we don't care about code quality for such builds anyway.
  */
 static inline void __scoped_seqlock_bug(void) { }
 #else