[v1] locking/lockdep: Enforce PROVE_RAW_LOCK_NESTING only if ARCH_SUPPORTS_RT

[PATCH] locking/lockdep: Enforce PROVE_RAW_LOCK_NESTING only if ARCH_SUPPORTS_RT

Posted by Waiman Long 1 year, 2 months ago

Relax the rule to set PROVE_RAW_LOCK_NESTING by default only for arches
that supports PREEMPT_RT.  For arches that do not support PREEMPT_RT,
they will not be forced to address irrelevant raw lock nesting issues
when they want to enable PROVE_LOCKING.

Signed-off-by: Waiman Long <longman@redhat.com>
---
 lib/Kconfig.debug | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug
index 1e37c62e8595..c53a498dc33f 100644
--- a/lib/Kconfig.debug
+++ b/lib/Kconfig.debug
@@ -1398,7 +1398,7 @@ config PROVE_LOCKING
 
 config PROVE_RAW_LOCK_NESTING
 	bool
-	depends on PROVE_LOCKING
+	depends on PROVE_LOCKING && ARCH_SUPPORTS_RT
 	default y
 	help
 	 Enable the raw_spinlock vs. spinlock nesting checks which ensure
-- 
2.47.0

Re: [PATCH] locking/lockdep: Enforce PROVE_RAW_LOCK_NESTING only if ARCH_SUPPORTS_RT

Posted by Sebastian Andrzej Siewior 1 year, 2 months ago

On 2024-11-26 18:11:54 [-0500], Waiman Long wrote:
> Relax the rule to set PROVE_RAW_LOCK_NESTING by default only for arches
> that supports PREEMPT_RT.  For arches that do not support PREEMPT_RT,
> they will not be forced to address irrelevant raw lock nesting issues
> when they want to enable PROVE_LOCKING.

I don't like the wording here. It is not "irrelevant raw lock nesting
issues". This is documented in Documentation/locking/locktypes.rst.
*IFF* we agree to ignore those because we don't want PREEMPT_RT on
certain architectures then okay. But please don't describe it as
irrelevant.

Sebastian

Re: [PATCH] locking/lockdep: Enforce PROVE_RAW_LOCK_NESTING only if ARCH_SUPPORTS_RT

Posted by Waiman Long 1 year, 2 months ago

On 11/27/24 11:44 AM, Sebastian Andrzej Siewior wrote:
> On 2024-11-26 18:11:54 [-0500], Waiman Long wrote:
>> Relax the rule to set PROVE_RAW_LOCK_NESTING by default only for arches
>> that supports PREEMPT_RT.  For arches that do not support PREEMPT_RT,
>> they will not be forced to address irrelevant raw lock nesting issues
>> when they want to enable PROVE_LOCKING.
> I don't like the wording here. It is not "irrelevant raw lock nesting
> issues". This is documented in Documentation/locking/locktypes.rst.
> *IFF* we agree to ignore those because we don't want PREEMPT_RT on
> certain architectures then okay. But please don't describe it as
> irrelevant.

Sorry for that. I will post a v2 patch to fix the wording.

Cheers,
Longman