Re: [PATCH v2 4/4] mm: thp: reparent the split queue during memcg offline

[Qi Zheng <zhengqi.arch@bytedance.com>]

Hi David,

On 9/24/25 8:38 PM, David Hildenbrand wrote:
> On 23.09.25 11:16, Qi Zheng wrote:
>> In the future, we will reparent LRU folios during memcg offline to
>> eliminate dying memory cgroups, which requires reparenting the split 
>> queue
>> to its parent.
>>
>> Similar to list_lru, the split queue is relatively independent and does
>> not need to be reparented along with objcg and LRU folios (holding
>> objcg lock and lru lock). So let's apply the same mechanism as list_lru
>> to reparent the split queue separately when memcg is offine.
>>
>> Signed-off-by: Qi Zheng <zhengqi.arch@bytedance.com>
>> ---
>>   include/linux/huge_mm.h |  2 ++
>>   include/linux/mmzone.h  |  1 +
>>   mm/huge_memory.c        | 39 +++++++++++++++++++++++++++++++++++++++
>>   mm/memcontrol.c         |  1 +
>>   mm/mm_init.c            |  1 +
>>   5 files changed, 44 insertions(+)
>>
>> diff --git a/include/linux/huge_mm.h b/include/linux/huge_mm.h
>> index f327d62fc9852..a0d4b751974d2 100644
>> --- a/include/linux/huge_mm.h
>> +++ b/include/linux/huge_mm.h
>> @@ -417,6 +417,7 @@ static inline int split_huge_page(struct page *page)
>>       return split_huge_page_to_list_to_order(page, NULL, ret);
>>   }
>>   void deferred_split_folio(struct folio *folio, bool partially_mapped);
>> +void reparent_deferred_split_queue(struct mem_cgroup *memcg);
>>   void __split_huge_pmd(struct vm_area_struct *vma, pmd_t *pmd,
>>           unsigned long address, bool freeze);
>> @@ -611,6 +612,7 @@ static inline int try_folio_split(struct folio 
>> *folio, struct page *page,
>>   }
>>   static inline void deferred_split_folio(struct folio *folio, bool 
>> partially_mapped) {}
>> +static inline void reparent_deferred_split_queue(struct mem_cgroup 
>> *memcg) {}
>>   #define split_huge_pmd(__vma, __pmd, __address)    \
>>       do { } while (0)
>> diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
>> index 7fb7331c57250..f3eb81fee056a 100644
>> --- a/include/linux/mmzone.h
>> +++ b/include/linux/mmzone.h
>> @@ -1346,6 +1346,7 @@ struct deferred_split {
>>       spinlock_t split_queue_lock;
>>       struct list_head split_queue;
>>       unsigned long split_queue_len;
>> +    bool is_dying;
> 
> It's a bit weird to query whether the "struct deferred_split" is dying. 
> Shouldn't this be a memcg property? (and in particular, not exist for 

There is indeed a CSS_DYING flag. But we must modify 'is_dying' under
the protection of the split_queue_lock, otherwise the folio may be added
back to the deferred_split of child memcg.

> the pglist_data part where it might not make sense at all?).

Maybe:

#ifdef CONFIG_MEMCG
     bool is_dying;
#endif

> 
>>   };
>>   #endif
>> diff --git a/mm/huge_memory.c b/mm/huge_memory.c
>> index 48b51e6230a67..de7806f759cba 100644
>> --- a/mm/huge_memory.c
>> +++ b/mm/huge_memory.c
>> @@ -1094,9 +1094,15 @@ static struct deferred_split 
>> *folio_split_queue_lock(struct folio *folio)
>>       struct deferred_split *queue;
>>       memcg = folio_memcg(folio);
>> +retry:
>>       queue = memcg ? &memcg->deferred_split_queue :
>>               &NODE_DATA(folio_nid(folio))->deferred_split_queue;
>>       spin_lock(&queue->split_queue_lock);
>> +    if (unlikely(queue->is_dying == true)) {
> 
> if (unlikely(queue->is_dying))

Will do.

> 
>> +        spin_unlock(&queue->split_queue_lock);
>> +        memcg = parent_mem_cgroup(memcg);
>> +        goto retry;
>> +    }
>>       return queue;
>>   }
>> @@ -1108,9 +1114,15 @@ folio_split_queue_lock_irqsave(struct folio 
>> *folio, unsigned long *flags)
>>       struct deferred_split *queue;
>>       memcg = folio_memcg(folio);
>> +retry:
>>       queue = memcg ? &memcg->deferred_split_queue :
>>               &NODE_DATA(folio_nid(folio))->deferred_split_queue;
>>       spin_lock_irqsave(&queue->split_queue_lock, *flags);
>> +    if (unlikely(queue->is_dying == true)) {
> 
> if (unlikely(queue->is_dying))

Will do.

> 
>> +        spin_unlock_irqrestore(&queue->split_queue_lock, *flags);
>> +        memcg = parent_mem_cgroup(memcg);
>> +        goto retry;
>> +    }
>>       return queue;
>>   }
> 
> Nothing else jumped at me, but I am not a memcg expert :)

Thanks,
Qi

>