Re: [PATCH] locking/rwsem: Fix kernel crash when spinning on RWSEM_OWNER_UNKNOWN

[Waiman Long]

On 1/15/20 1:50 AM, Christoph Hellwig wrote:
> On Tue, Jan 14, 2020 at 02:03:03PM -0500, Waiman Long wrote:
>> The commit 91d2a812dfb9 ("locking/rwsem: Make handoff writer
>> optimistically spin on owner") will allow a recently woken up waiting
>> writer to spin on the owner. Unfortunately, if the owner happens to be
>> RWSEM_OWNER_UNKNOWN, the code will incorrectly spin on it leading to a
>> kernel crash. This is fixed by passing the proper non-spinnable bits
>> to rwsem_spin_on_owner() so that RWSEM_OWNER_UNKNOWN will be treated
>> as a non-spinnable target.
>>
>> Fixes: 91d2a812dfb9 ("locking/rwsem: Make handoff writer optimistically spin on owner")
>>
>> Reported-by: Christoph Hellwig <hch@xxxxxx>
>> Signed-off-by: Waiman Long <longman@xxxxxxxxxx>
> This survives all the tests that showed the problems with the original
> code:
>
> Tested-by: Christoph Hellwig <hch@xxxxxx>
>
Thanks for the testing.
>>  		if ((wstate == WRITER_HANDOFF) &&
>> -		    (rwsem_spin_on_owner(sem, 0) == OWNER_NULL))
>> +		    rwsem_spin_on_owner(sem, RWSEM_NONSPINNABLE) == OWNER_NULL)
> Nit: the inner braces in the first half of the conditional aren't required
> either.

Yes, it is inconsistent and so is not good. I will post a v2 patch to
fix that.

Cheers,
Longman