Re: [PATCH v2] Reorder some fields in struct rq.

[Madadi Vineeth Reddy]

Hi Blake,

On 31/07/25 02:26, Blake Jones wrote:
> This colocates some hot fields in "struct rq" to be on the same cache line
> as others that are often accessed at the same time or in similar ways.
> 

[..snip..]

> 
> This patch does not change the size of "struct rq" on machines with 64-byte
> cache lines. The additional "____cacheline_aligned" to put the runqueue
> lock on the next cache line will add an additional 64 bytes of padding on
> machines with 128-byte cache lines; although this is unfortunate, it seemed
> more likely to lead to stably good performance than e.g. by just putting
> the runqueue lock somewhere in the middle of the structure and hoping it
> wasn't on an otherwise busy cache line.

This change introduced an 88 byte hole due to having __lock in a different
cache line on Power11 which is 128 byte architecture which led to one cacheline
more than before.

Tested with your custom test case (thanks for sharing) and observed around
~5% decrease in the number of cycles, along with a slight increase in user
time — both are positive indicators.

Also ran ebizzy, which doesn’t seem to be impacted. I think it would be good
to run a set of standard benchmarks like schbench, ebizzy, hackbench, and
stress-ng, along with a real-life workload, to ensure there’s no negative
impact. I saw that hackbench was tried, but including those numbers would
be helpful.

Reviewed-by: Madadi Vineeth Reddy <vineethr@xxxxxxxxxxxxx>
Tested-by: Madadi Vineeth Reddy <vineethr@xxxxxxxxxxxxx>

Thanks,
Madadi Vineeth Reddy

> 
> I ran "hackbench" to test this change, but it didn't show very conclusive
> results.  Looking at a profile of the hackbench run, it was spending 95% of
> its cycles inside __alloc_skb(), __kfree_skb(), or kmem_cache_free() -
> almost all of which was spent updating memcg counters or contending on the
> list_lock in kmem_cache_node. In contrast, it spent less than 0.5% of its
> cycles inside either schedule() or try_to_wake_up(). So it's not surprising
> that it didn't show useful results here.
> 

[..snip..]

> @@ -1182,8 +1199,6 @@ struct rq {
>  	struct root_domain		*rd;
>  	struct sched_domain __rcu	*sd;
>  
> -	unsigned long		cpu_capacity;
> -
>  	struct balance_callback *balance_callback;
>  
>  	unsigned char		nohz_idle_balance;