Re: [PATCH 1/6] sched/numa, mm: make numa migrate functions to take a folio

[Kefeng Wang]

On 2023/9/20 11:05, Huang, Ying wrote:
> Kefeng Wang <wangkefeng.wang@xxxxxxxxxx> writes:
>
> > The cpuid(or access time) is stored in the head page for THP, so it is
>
> s/cpuid/cpupid/

Will fix.

> > safely to make should_numa_migrate_memory() and numa_hint_fault_latency()
> > to take a folio. This is in preparation for large folio numa balancing.
> >
> > Signed-off-by: Kefeng Wang <wangkefeng.wang@xxxxxxxxxx>
> > ---
> >   include/linux/sched/numa_balancing.h |  4 ++--
> >   kernel/sched/fair.c                  | 12 ++++++------
> >   mm/mempolicy.c                       |  3 ++-
> >   3 files changed, 10 insertions(+), 9 deletions(-)
> >
> > diff --git a/include/linux/sched/numa_balancing.h b/include/linux/sched/numa_balancing.h
> > index 3988762efe15..a38528c28665 100644
> > --- a/include/linux/sched/numa_balancing.h
> > +++ b/include/linux/sched/numa_balancing.h
> > @@ -20,7 +20,7 @@ extern void task_numa_fault(int last_node, int node, int pages, int flags);
> >   extern pid_t task_numa_group_id(struct task_struct *p);
> >   extern void set_numabalancing_state(bool enabled);
> >   extern void task_numa_free(struct task_struct *p, bool final);
> > -extern bool should_numa_migrate_memory(struct task_struct *p, struct page *page,
> > +extern bool should_numa_migrate_memory(struct task_struct *p, struct folio *folio,
> >   					int src_nid, int dst_cpu);
> >   #else
> >   static inline void task_numa_fault(int last_node, int node, int pages,
> > @@ -38,7 +38,7 @@ static inline void task_numa_free(struct task_struct *p, bool final)
> >   {
> >   }
> >   static inline bool should_numa_migrate_memory(struct task_struct *p,
> > -				struct page *page, int src_nid, int dst_cpu)
> > +				struct folio *folio, int src_nid, int dst_cpu)
> >   {
> >   	return true;
> >   }
> > diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> > index cb225921bbca..683cc1e417d7 100644
> > --- a/kernel/sched/fair.c
> > +++ b/kernel/sched/fair.c
> > @@ -1722,12 +1722,12 @@ static bool pgdat_free_space_enough(struct pglist_data *pgdat)
> >    * The smaller the hint page fault latency, the higher the possibility
> >    * for the page to be hot.
> >    */
> > -static int numa_hint_fault_latency(struct page *page)
> > +static int numa_hint_fault_latency(struct folio *folio)
> >   {
> >   	int last_time, time;
> >   
> >   	time = jiffies_to_msecs(jiffies);
> > -	last_time = xchg_page_access_time(page, time);
> > +	last_time = xchg_page_access_time(&folio->page, time);
>
> How about define xchg_folio_access_time() and folio_cpupid_xchg_last()?
Yes, like 
page_cpupid_last()/xchg_page_access_time()/page_cpupid_xchg_last(),
we could do it later to change the caller to use a folio, and rename them.


> --
> Best Regards,
> Huang, Ying