Re: [PATCH 3/3] tools: memory-model: Improve data-race detection

From: Akira Yokosawa
Date: Sun Jun 23 2019 - 05:37:37 EST


Hi Paul and Alan,

On 2019/06/22 8:54, Paul E. McKenney wrote:
> On Fri, Jun 21, 2019 at 10:25:23AM -0400, Alan Stern wrote:
>> On Fri, 21 Jun 2019, Andrea Parri wrote:
>>
>>> On Thu, Jun 20, 2019 at 11:55:58AM -0400, Alan Stern wrote:
>>>> Herbert Xu recently reported a problem concerning RCU and compiler
>>>> barriers. In the course of discussing the problem, he put forth a
>>>> litmus test which illustrated a serious defect in the Linux Kernel
>>>> Memory Model's data-race-detection code.

I was not involved in the mail thread and wondering what the litmus test
looked like. Some searching of the archive has suggested that Alan presented
a properly formatted test based on Herbert's idea in [1].

[1]: https://lore.kernel.org/lkml/Pine.LNX.4.44L0.1906041026570.1731-100000@xxxxxxxxxxxxxxxxxxxx/

If this is the case, adding the link (or message id) in the change
log would help people see the circumstances, I suppose.
Paul, can you amend the change log?

I ran herd7 on said litmus test at both "lkmm" and "dev" of -rcu and
confirmed that this patch fixes the result.

So,

Tested-by: Akira Yokosawa <akiyks@xxxxxxxxx>

Thanks, Akira

>>>>
>>>> The defect was that the LKMM assumed visibility and executes-before
>>>> ordering of plain accesses had to be mediated by marked accesses. In
>>>> Herbert's litmus test this wasn't so, and the LKMM claimed the litmus
>>>> test was allowed and contained a data race although neither is true.
>>>>
>>>> In fact, plain accesses can be ordered by fences even in the absence
>>>> of marked accesses. In most cases this doesn't matter, because most
>>>> fences only order accesses within a single thread. But the rcu-fence
>>>> relation is different; it can order (and induce visibility between)
>>>> accesses in different threads -- events which otherwise might be
>>>> concurrent. This makes it relevant to data-race detection.
>>>>
>>>> This patch makes two changes to the memory model to incorporate the
>>>> new insight:
>>>>
>>>> If a store is separated by a fence from another access,
>>>> the store is necessarily visible to the other access (as
>>>> reflected in the ww-vis and wr-vis relations). Similarly,
>>>> if a load is separated by a fence from another access then
>>>> the load necessarily executes before the other access (as
>>>> reflected in the rw-xbstar relation).
>>>>
>>>> If a store is separated by a strong fence from a marked access
>>>> then it is necessarily visible to any access that executes
>>>> after the marked access (as reflected in the ww-vis and wr-vis
>>>> relations).
>>>>
>>>> With these changes, the LKMM gives the desired result for Herbert's
>>>> litmus test and other related ones.
>>>>
>>>> Signed-off-by: Alan Stern <stern@xxxxxxxxxxxxxxxxxxx>
>>>> Reported-by: Herbert Xu <herbert@xxxxxxxxxxxxxxxxxxx>
>>>
>>> For the entire series:
>>>
>>> Acked-by: Andrea Parri <andrea.parri@xxxxxxxxxxxxxxxxxxxx>
>>>
>>> Two nits, but up to Paul AFAIAC:
>>>
>>> - This is a first time for "tools: memory-model:" in Subject; we were
>>> kind of converging to "tools/memory-model:"...
>>
>> Yeah, sure. That's the sort of detail I have a hard time remembering.
>>
>>> - The report preceded the patch; we might as well reflect this in the
>>> order of the tags.
>>
>> Either way is okay with me.
>
> I applied Andrea's acks and edited as called out above, thank you both!
>
> Thanx, Paul
>