Re: [PATCH v2] rcu: Reduce synchronize_rcu() delays when all wait heads are in use

[Neeraj Upadhyay <Neeraj.Upadhyay@amd.com>]

On 4/5/2024 3:12 AM, Frederic Weisbecker wrote:
> Le Wed, Apr 03, 2024 at 04:22:12PM +0530, Neeraj Upadhyay a écrit :
>> When all wait heads are in use, which can happen when
>> rcu_sr_normal_gp_cleanup_work()'s callback processing
>> is slow, any new synchronize_rcu() user's rcu_synchronize
>> node's processing is deferred to future GP periods. This
>> can result in long list of synchronize_rcu() invocations
>> waiting for full grace period processing, which can delay
>> freeing of memory. Mitigate this problem by using first
>> node in the list as wait tail when all wait heads are in use.
>> While methods to speed up callback processing would be needed
>> to recover from this situation, allowing new nodes to complete
>> their grace period can help prevent delays due to a fixed
>> number of wait head nodes.
>>
>> Signed-off-by: Neeraj Upadhyay <Neeraj.Upadhyay@amd.com>
> 
> Looking at it again, I'm not sure if it's a good idea to
> optimize the thing that far. It's already a tricky state machine
> to review and the workqueue has SR_NORMAL_GP_WAIT_HEAD_MAX - 1 = 4
> grace periods worth of time to execute. Such a tense situation may
> happen of course but, should we really work around that?
> 
> I let you guys judge. In the meantime, I haven't found correctness

I agree that this adds more complexity for handling a scenario
which is not expected to happen often. Also, this does not help
much to recover from the situation, as most of the callbacks are still
blocked on kworker execution. Intent was to keep the patch ready, in
case we see fixed SR_NORMAL_GP_WAIT_HEAD_MAX  as a blocking factor.
It's fine from my side if we want to hold off this one. Uladzislau
what do you think?

> issues:
> 
> Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
> 

Thanks!


- Neeraj

> Thanks.