Re: [PATCH v5] locking/rwsem: Make handoff bit handling more consistent

Re: [PATCH v5] locking/rwsem: Make handoff bit handling more consistent

[Waiman Long]

On 4/12/22 12:28, john.p.donnelly@oracle.com wrote:
> On 4/11/22 4:07 PM, Waiman Long wrote:
>>
>> On 4/11/22 17:03, john.p.donnelly@oracle.com wrote:
>>>
>>>>>
>>>>> I have reached out to Waiman and he suggested this for our next 
>>>>> test pass:
>>>>>
>>>>>
>>>>> 1ee326196c6658 locking/rwsem: Always try to wake waiters in 
>>>>> out_nolock path
>>>>
>>>> Does this commit help to avoid the lockup problem?
>>>>
>>>> Commit 1ee326196c6658 fixes a potential missed wakeup problem when 
>>>> a reader first in the wait queue is interrupted out without 
>>>> acquiring the lock. It is actually not a fix for commit 
>>>> d257cc8cb8d5. However, this commit changes the out_nolock path 
>>>> behavior of writers by leaving the handoff bit set when the wait 
>>>> queue isn't empty. That likely makes the missed wakeup problem 
>>>> easier to reproduce.
>>>>
>>>> Cheers,
>>>> Longman
>>>>
>>>
>>> Hi,
>>>
>>>
>>> We are testing now
>>>
>>> ETA for fio soak test completion is  ~15hr from now.
>>>
>>> I wanted to share the stack traces for future reference + occurrences.
>>>
>> I am looking forward to your testing results tomorrow.
>>
>> Cheers,
>> Longman
>>
> Hi
>
>  Our 24hr fio soak test with :
>
>  1ee326196c6658 locking/rwsem: Always try to wake waiters in 
> out_nolock path
>
>
>  applied to 5.15.30  passed.
>
>  I suggest you append  1ee326196c6658 with :
>
>
>  cc: stable
>
>   Fixes: d257cc8cb8d5 ("locking/rwsem: Make handoff bit handling more 
> consistent")
>
>
> I'll leave the implementation details up to the core maintainers how 
> to do that ;-)

Thanks for the test.

The patch has already been in the tip tree. It may not be easy to add a 
Fixes tag to it. Anyway, I will encourage stable tree maintainer to take 
it as it does fix a problem as shown in your test.

Cheers,
Longman

Re: [PATCH v5] locking/rwsem: Make handoff bit handling more consistent

[Greg KH]

On Tue, Apr 12, 2022 at 01:04:05PM -0400, Waiman Long wrote:
> The patch has already been in the tip tree. It may not be easy to add a
> Fixes tag to it. Anyway, I will encourage stable tree maintainer to take it
> as it does fix a problem as shown in your test.

I have no idea what you want me to do here.  Please be explicit...

thanks,

greg k-h

Re: [PATCH v5] locking/rwsem: Make handoff bit handling more consistent

[Waiman Long]

On 4/14/22 06:48, Greg KH wrote:
> On Tue, Apr 12, 2022 at 01:04:05PM -0400, Waiman Long wrote:
>> The patch has already been in the tip tree. It may not be easy to add a
>> Fixes tag to it. Anyway, I will encourage stable tree maintainer to take it
>> as it does fix a problem as shown in your test.
> I have no idea what you want me to do here.  Please be explicit...
>
I would like the stable trees to include commit 1ee326196c66 
("locking/rwsem: Always try to wake waiters in out_nolock path") after 
it is merged into the mainline in the v5.19 merge window. It should be 
applied to the stable trees that have incorporated the backport of 
commit d257cc8cb8d5 ("locking/rwsem: Make handoff bit handling more 
consistent") since this commit seems to make the missed wakeup problem 
easier to show up.

Cheers,
Longman

Re: [PATCH v5] locking/rwsem: Make handoff bit handling more consistent

[Greg KH]

On Thu, Apr 14, 2022 at 11:18:23AM -0400, Waiman Long wrote:
> On 4/14/22 06:48, Greg KH wrote:
> > On Tue, Apr 12, 2022 at 01:04:05PM -0400, Waiman Long wrote:
> > > The patch has already been in the tip tree. It may not be easy to add a
> > > Fixes tag to it. Anyway, I will encourage stable tree maintainer to take it
> > > as it does fix a problem as shown in your test.
> > I have no idea what you want me to do here.  Please be explicit...
> > 
> I would like the stable trees to include commit 1ee326196c66
> ("locking/rwsem: Always try to wake waiters in out_nolock path") after it is
> merged into the mainline in the v5.19 merge window. It should be applied to
> the stable trees that have incorporated the backport of commit d257cc8cb8d5
> ("locking/rwsem: Make handoff bit handling more consistent") since this
> commit seems to make the missed wakeup problem easier to show up.

Please let us know when this is merged into Linus's tree.  Nothing we
can do until then and I'm not going to remember this in a few months
anyway.

thanks,

greg k-h

Re: [PATCH v5] locking/rwsem: Make handoff bit handling more consistent

[Waiman Long]

On 4/14/22 11:42, Greg KH wrote:
> On Thu, Apr 14, 2022 at 11:18:23AM -0400, Waiman Long wrote:
>> On 4/14/22 06:48, Greg KH wrote:
>>> On Tue, Apr 12, 2022 at 01:04:05PM -0400, Waiman Long wrote:
>>>> The patch has already been in the tip tree. It may not be easy to add a
>>>> Fixes tag to it. Anyway, I will encourage stable tree maintainer to take it
>>>> as it does fix a problem as shown in your test.
>>> I have no idea what you want me to do here.  Please be explicit...
>>>
>> I would like the stable trees to include commit 1ee326196c66
>> ("locking/rwsem: Always try to wake waiters in out_nolock path") after it is
>> merged into the mainline in the v5.19 merge window. It should be applied to
>> the stable trees that have incorporated the backport of commit d257cc8cb8d5
>> ("locking/rwsem: Make handoff bit handling more consistent") since this
>> commit seems to make the missed wakeup problem easier to show up.
> Please let us know when this is merged into Linus's tree.  Nothing we
> can do until then and I'm not going to remember this in a few months
> anyway.

Sure, I will remind you again after the merge happens.

Cheers,
Longman