[PATCH] locking/rtmutex: Add the missing 'ES' of macro CONFIG_DEBUG_RT_MUTEXES

[PATCH] locking/rtmutex: Add the missing 'ES' of macro CONFIG_DEBUG_RT_MUTEXES

[Zhen Lei]

The build option name is defined as DEBUG_RT_MUTEXES in lib/Kconfig.debug,
commit f41dcc18698e ("locking/rtmutex: Move debug functions as inlines
into common header") can also corroborate this.

Fixes: f7efc4799f81 ("locking/rtmutex: Inline chainwalk depth check")
Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com>
---
 kernel/locking/rtmutex.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/kernel/locking/rtmutex.c b/kernel/locking/rtmutex.c
index b5d9bb5202c6..ad0db322ed3b 100644
--- a/kernel/locking/rtmutex.c
+++ b/kernel/locking/rtmutex.c
@@ -343,7 +343,7 @@ static __always_inline bool
 rt_mutex_cond_detect_deadlock(struct rt_mutex_waiter *waiter,
 			      enum rtmutex_chainwalk chwalk)
 {
-	if (IS_ENABLED(CONFIG_DEBUG_RT_MUTEX))
+	if (IS_ENABLED(CONFIG_DEBUG_RT_MUTEXES))
 		return waiter != NULL;
 	return chwalk == RT_MUTEX_FULL_CHAINWALK;
 }
-- 
2.25.1

Re: [PATCH] locking/rtmutex: Add the missing 'ES' of macro CONFIG_DEBUG_RT_MUTEXES

[Will Deacon]

On Sat, Jul 31, 2021 at 08:30:11PM +0800, Zhen Lei wrote:
> The build option name is defined as DEBUG_RT_MUTEXES in lib/Kconfig.debug,
> commit f41dcc18698e ("locking/rtmutex: Move debug functions as inlines
> into common header") can also corroborate this.
> 
> Fixes: f7efc4799f81 ("locking/rtmutex: Inline chainwalk depth check")
> Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com>
> ---
>  kernel/locking/rtmutex.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/kernel/locking/rtmutex.c b/kernel/locking/rtmutex.c
> index b5d9bb5202c6..ad0db322ed3b 100644
> --- a/kernel/locking/rtmutex.c
> +++ b/kernel/locking/rtmutex.c
> @@ -343,7 +343,7 @@ static __always_inline bool
>  rt_mutex_cond_detect_deadlock(struct rt_mutex_waiter *waiter,
>  			      enum rtmutex_chainwalk chwalk)
>  {
> -	if (IS_ENABLED(CONFIG_DEBUG_RT_MUTEX))
> +	if (IS_ENABLED(CONFIG_DEBUG_RT_MUTEXES))
>  		return waiter != NULL;
>  	return chwalk == RT_MUTEX_FULL_CHAINWALK;

Oops, yes. How did you find this?

Acked-by: Will Deacon <will@kernel.org>

Will

Re: [PATCH] locking/rtmutex: Add the missing 'ES' of macro CONFIG_DEBUG_RT_MUTEXES

[Leizhen (ThunderTown)]

On 2021/8/9 22:41, Will Deacon wrote:
> On Sat, Jul 31, 2021 at 08:30:11PM +0800, Zhen Lei wrote:
>> The build option name is defined as DEBUG_RT_MUTEXES in lib/Kconfig.debug,
>> commit f41dcc18698e ("locking/rtmutex: Move debug functions as inlines
>> into common header") can also corroborate this.
>>
>> Fixes: f7efc4799f81 ("locking/rtmutex: Inline chainwalk depth check")
>> Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com>
>> ---
>>  kernel/locking/rtmutex.c | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/kernel/locking/rtmutex.c b/kernel/locking/rtmutex.c
>> index b5d9bb5202c6..ad0db322ed3b 100644
>> --- a/kernel/locking/rtmutex.c
>> +++ b/kernel/locking/rtmutex.c
>> @@ -343,7 +343,7 @@ static __always_inline bool
>>  rt_mutex_cond_detect_deadlock(struct rt_mutex_waiter *waiter,
>>  			      enum rtmutex_chainwalk chwalk)
>>  {
>> -	if (IS_ENABLED(CONFIG_DEBUG_RT_MUTEX))
>> +	if (IS_ENABLED(CONFIG_DEBUG_RT_MUTEXES))
>>  		return waiter != NULL;
>>  	return chwalk == RT_MUTEX_FULL_CHAINWALK;
> 
> Oops, yes. How did you find this?

I found it by chance. When locating a futex problem, I was confused when
reading this part of the code. I used "git grep -wn CONFIG_DEBUG_RT_MUTEX"
to search for the implementation in other places. But I found that it was
not used in other places, then I checked the patch history and found it.

> 
> Acked-by: Will Deacon <will@kernel.org>
> 
> Will
> .
> 

Re: [PATCH] locking/rtmutex: Add the missing 'ES' of macro CONFIG_DEBUG_RT_MUTEXES

[Boqun Feng]

On Tue, Aug 10, 2021 at 09:21:15AM +0800, Leizhen (ThunderTown) wrote:
> 
> 
> On 2021/8/9 22:41, Will Deacon wrote:
> > On Sat, Jul 31, 2021 at 08:30:11PM +0800, Zhen Lei wrote:
> >> The build option name is defined as DEBUG_RT_MUTEXES in lib/Kconfig.debug,
> >> commit f41dcc18698e ("locking/rtmutex: Move debug functions as inlines
> >> into common header") can also corroborate this.
> >>
> >> Fixes: f7efc4799f81 ("locking/rtmutex: Inline chainwalk depth check")
> >> Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com>
> >> ---
> >>  kernel/locking/rtmutex.c | 2 +-
> >>  1 file changed, 1 insertion(+), 1 deletion(-)
> >>
> >> diff --git a/kernel/locking/rtmutex.c b/kernel/locking/rtmutex.c
> >> index b5d9bb5202c6..ad0db322ed3b 100644
> >> --- a/kernel/locking/rtmutex.c
> >> +++ b/kernel/locking/rtmutex.c
> >> @@ -343,7 +343,7 @@ static __always_inline bool
> >>  rt_mutex_cond_detect_deadlock(struct rt_mutex_waiter *waiter,
> >>  			      enum rtmutex_chainwalk chwalk)
> >>  {
> >> -	if (IS_ENABLED(CONFIG_DEBUG_RT_MUTEX))
> >> +	if (IS_ENABLED(CONFIG_DEBUG_RT_MUTEXES))
> >>  		return waiter != NULL;
> >>  	return chwalk == RT_MUTEX_FULL_CHAINWALK;
> > 
> > Oops, yes. How did you find this?
> 
> I found it by chance. When locating a futex problem, I was confused when
> reading this part of the code. I used "git grep -wn CONFIG_DEBUG_RT_MUTEX"
> to search for the implementation in other places. But I found that it was
> not used in other places, then I checked the patch history and found it.
> 

FWIW, script/checkkconfigsymbols.py can be used to detect usage of
undefined CONFIG_*, Zhouyi Zhou used it to find a few cases in RCU
recently:

	https://lore.kernel.org/lkml/CAABZP2wuWtGAGRqWJb3Gewm5VLZdZ_C=LRZsFbaG3jcQabO3qA@mail.gmail.com/

And thanks for the fix:

Acked-by: Boqun Feng <boqun.feng@gmail.com>

Regards,
Boqun

> > 
> > Acked-by: Will Deacon <will@kernel.org>
> > 
> > Will
> > .
> > 

[tip: locking/urgent] locking/rtmutex: Use the correct rtmutex debugging config option

[tip-bot2 for Zhen Lei]

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

Commit-ID:     07d25971b220e477eb019fcb520a9f2e3ac966af
Gitweb:        https://git.kernel.org/tip/07d25971b220e477eb019fcb520a9f2e3ac966af
Author:        Zhen Lei <thunder.leizhen@huawei.com>
AuthorDate:    Sat, 31 Jul 2021 20:30:11 +08:00
Committer:     Thomas Gleixner <tglx@linutronix.de>
CommitterDate: Tue, 10 Aug 2021 08:21:52 +02:00

locking/rtmutex: Use the correct rtmutex debugging config option

It's CONFIG_DEBUG_RT_MUTEXES not CONFIG_DEBUG_RT_MUTEX.

Fixes: f7efc4799f81 ("locking/rtmutex: Inline chainwalk depth check")
Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Will Deacon <will@kernel.org>
Acked-by: Boqun Feng <boqun.feng@gmail.com>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/20210731123011.4555-1-thunder.leizhen@huawei.com

---
 kernel/locking/rtmutex.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/kernel/locking/rtmutex.c b/kernel/locking/rtmutex.c
index b5d9bb5..ad0db32 100644
--- a/kernel/locking/rtmutex.c
+++ b/kernel/locking/rtmutex.c
@@ -343,7 +343,7 @@ static __always_inline bool
 rt_mutex_cond_detect_deadlock(struct rt_mutex_waiter *waiter,
 			      enum rtmutex_chainwalk chwalk)
 {
-	if (IS_ENABLED(CONFIG_DEBUG_RT_MUTEX))
+	if (IS_ENABLED(CONFIG_DEBUG_RT_MUTEXES))
 		return waiter != NULL;
 	return chwalk == RT_MUTEX_FULL_CHAINWALK;
 }