Re: [PATCH v2] locking/rwsem: Add DEBUG_RWSEMS to look for lock/unlock mismatches
From: Waiman Long
Date: Tue Mar 27 2018 - 16:37:41 EST
On 03/27/2018 03:56 PM, Ingo Molnar wrote:
> * Waiman Long <longman@xxxxxxxxxx> wrote:
>
>> On 03/27/2018 02:06 AM, Ingo Molnar wrote:
>>> * Waiman Long <longman@xxxxxxxxxx> wrote:
>>>
>>>> For a rwsem, locking can either be exclusive or shared. The corresponding
>>>> exclusive or shared unlock must be used. Otherwise, the protected data
>>>> structures may get corrupted or the lock may be in an inconsistent state.
>>>>
>>>> In order to detect such anomaly, a new configuration option DEBUG_RWSEMS
>>>> is added which can be enabled to look for such mismatches and print
>>>> warnings that that happens.
>>>> diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug
>>>> index 64155e3..0958192 100644
>>>> --- a/lib/Kconfig.debug
>>>> +++ b/lib/Kconfig.debug
>>>> @@ -1075,6 +1075,13 @@ config DEBUG_WW_MUTEX_SLOWPATH
>>>> even a debug kernel. If you are a driver writer, enable it. If
>>>> you are a distro, do not.
>>>>
>>>> +config DEBUG_RWSEMS
>>>> + bool "RW Semaphore debugging: basic checks"
>>>> + depends on DEBUG_KERNEL && RWSEM_SPIN_ON_OWNER
>>>> + help
>>>> + This feature allows mismatched rw semaphore locks and unlocks
>>>> + to be detected and reported.
>>>> +
>>> Makes sense - but this should also be integrated into the rest of lock debugging
>>> Kconfig hierarchy similar to DEBUG_MUTEXES: i.e. DEBUG_LOCK_ALLOC, PROVE_LOCKING,
>>> etc. should select this new lock debugging option as well.
>>>
>>> People generally are not supposed to know and configure the finer details,
>>> CONFIG_LOCK_DEBUGGING=y is a one-stop-shop in this regard.
>> I can add a patch to rework lock debugging configuration code. Do you
>> want to hide the individual config options but still allow them to be
>> enabled manually? Alternatively we can also just add a master lock
>> debugging option to select them all.
> They should work like mutex debugging: it's auto-selected by PROVE_LOCKING, but
> can be enabled individually as well if PROVE_LOCKING is disabled.
Got it.
Thanks,
Longman