On Wed, Oct 26, 2022 at 04:26:28PM +0800, Yin, Fengwei wrote:
> Hi Feng,
>
> On 10/26/2022 3:43 PM, Feng Tang wrote:
> > In page reclaim path, memory could be demoted from faster memory tier
> > to slower memory tier. Currently, there is no check about cpuset's
> > memory policy, that even if the target demotion node is not allowd
> > by cpuset, the demotion will still happen, which breaks the cpuset
> > semantics.
> >
> > So add cpuset policy check in the demotion path and skip demotion
> > if the demotion targets are not allowed by cpuset.
> >
> > Signed-off-by: Feng Tang <feng.tang@xxxxxxxxx>
> > ---
> > Hi reviewers,
> >
> > For easy bisectable, I combined the cpuset change and mm change
> > in one patch, if you prefer to separate them, I can turn it into
> > 2 patches.
> >
> > Thanks,
> > Feng
> >
> > include/linux/cpuset.h | 6 ++++++
> > kernel/cgroup/cpuset.c | 29 +++++++++++++++++++++++++++++
> > mm/vmscan.c | 35 ++++++++++++++++++++++++++++++++---
> > 3 files changed, 67 insertions(+), 3 deletions(-)
> >
> > diff --git a/include/linux/cpuset.h b/include/linux/cpuset.h
> > index d58e0476ee8e..6fcce2bd2631 100644
> > --- a/include/linux/cpuset.h
> > +++ b/include/linux/cpuset.h
> > @@ -178,6 +178,8 @@ static inline void set_mems_allowed(nodemask_t nodemask)
> > task_unlock(current);
> > }
> >
> > +extern void cpuset_get_allowed_mem_nodes(struct cgroup *cgroup,
> > + nodemask_t *nmask);
> > #else /* !CONFIG_CPUSETS */
> >
> > static inline bool cpusets_enabled(void) { return false; }
> > @@ -299,6 +301,10 @@ static inline bool read_mems_allowed_retry(unsigned int seq)
> > return false;
> > }
> >
> > +static inline void cpuset_get_allowed_mem_nodes(struct cgroup *cgroup,
> > + nodemask_t *nmask)
> > +{
> > +}
> > #endif /* !CONFIG_CPUSETS */
> >
> > #endif /* _LINUX_CPUSET_H */
> > diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c
> > index 3ea2e836e93e..cbb118c0502f 100644
> > --- a/kernel/cgroup/cpuset.c
> > +++ b/kernel/cgroup/cpuset.c
> > @@ -3750,6 +3750,35 @@ nodemask_t cpuset_mems_allowed(struct task_struct *tsk)
> > return mask;
> > }
> >
> > +/*
> > + * Retrieve the allowed memory nodemask for a cgroup.
> > + *
> > + * Set *nmask to cpuset's effective allowed nodemask for cgroup v2,
> > + * and NODE_MASK_ALL (means no constraint) for cgroup v1 where there
> > + * is no guaranteed association from a cgroup to a cpuset.
> > + */
> > +void cpuset_get_allowed_mem_nodes(struct cgroup *cgroup, nodemask_t *nmask)
> > +{
> > + struct cgroup_subsys_state *css;
> > + struct cpuset *cs;
> > +
> > + if (!is_in_v2_mode()) {
> > + *nmask = NODE_MASK_ALL;
> > + return;
> > + }
> > +
> > + rcu_read_lock();
> > + css = cgroup_e_css(cgroup, &cpuset_cgrp_subsys);
> > + if (css) {
> > + css_get(css);
> > + cs = css_cs(css);
> > + *nmask = cs->effective_mems;
> > + css_put(css);
> > + }
> > +
> > + rcu_read_unlock();
> > +}
> > +
> > /**
> > * cpuset_nodemask_valid_mems_allowed - check nodemask vs. current mems_allowed
> > * @nodemask: the nodemask to be checked
> > diff --git a/mm/vmscan.c b/mm/vmscan.c
> > index 18f6497994ec..c205d98283bc 100644
> > --- a/mm/vmscan.c
> > +++ b/mm/vmscan.c
> > @@ -1537,9 +1537,21 @@ static struct page *alloc_demote_page(struct page *page, unsigned long private)
> > {
> > struct page *target_page;
> > nodemask_t *allowed_mask;
> > - struct migration_target_control *mtc;
> > + struct migration_target_control *mtc = (void *)private;
> >
> > - mtc = (struct migration_target_control *)private;
> > +#if IS_ENABLED(CONFIG_MEMCG) && IS_ENABLED(CONFIG_CPUSETS)
> > + struct mem_cgroup *memcg;
> > + nodemask_t cpuset_nmask;
> > +
> > + memcg = page_memcg(page);
> > + cpuset_get_allowed_mem_nodes(memcg->css.cgroup, &cpuset_nmask);
> > +
> > + if (!node_isset(mtc->nid, cpuset_nmask)) {
> > + if (mtc->nmask)
> > + nodes_and(*mtc->nmask, *mtc->nmask, cpuset_nmask);
> > + return alloc_migration_target(page, (unsigned long)mtc);
> > + }
> > +#endif
> >
> > allowed_mask = mtc->nmask;
> > /*
> > @@ -1649,6 +1661,7 @@ static unsigned int shrink_folio_list(struct list_head *folio_list,
> > enum folio_references references = FOLIOREF_RECLAIM;
> > bool dirty, writeback;
> > unsigned int nr_pages;
> > + bool skip_this_demotion = false;
> >
> > cond_resched();
> >
> > @@ -1658,6 +1671,22 @@ static unsigned int shrink_folio_list(struct list_head *folio_list,
> > if (!folio_trylock(folio))
> > goto keep;
> >
> > +#if IS_ENABLED(CONFIG_MEMCG) && IS_ENABLED(CONFIG_CPUSETS)
> > + if (do_demote_pass) {
> > + struct mem_cgroup *memcg;
> > + nodemask_t nmask, nmask1;
> > +
> > + node_get_allowed_targets(pgdat, &nmask);
> > + memcg = folio_memcg(folio);
> > + if (memcg)
> Doesn't check memcg in the change to alloc_demote_page(). What's the difference here?
>
> > + cpuset_get_allowed_mem_nodes(memcg->css.cgroup,
> > + &nmask1);
> > +
> > + if (!nodes_intersects(nmask, nmask1))
> If memcg is NULL, nmask1 will have an uninitialized value. Thanks.
Good catch! Yes, it should be initialized to NODE_MASK_ALL.
Actually I was not sure if I need to check 'memcg == NULL' case, while
I was under the impression that for page on LRU list, it's memcg is
either a specific memcg or the 'root_mem_cgroup'. I will double
check that.
Thanks,
Feng
> Regards
> Yin, Fengwei
>
> > + skip_this_demotion = true;
> > + }
> > +#endif
> > +
> > VM_BUG_ON_FOLIO(folio_test_active(folio), folio);
> >
> > nr_pages = folio_nr_pages(folio);
> > @@ -1799,7 +1828,7 @@ static unsigned int shrink_folio_list(struct list_head *folio_list,
> > * Before reclaiming the folio, try to relocate
> > * its contents to another node.
> > */
> > - if (do_demote_pass &&
> > + if (do_demote_pass && !skip_this_demotion &&
> > (thp_migration_supported() || !folio_test_large(folio))) {
> > list_add(&folio->lru, &demote_folios);
> > folio_unlock(folio);
