Re: [PATCH 1/2] mm/memcg: use kmem_cache when alloc memcg

[Huan Yang <link@vivo.com>]

在 2025/4/24 05:59, Andrew Morton 写道:
> On Wed, 23 Apr 2025 16:43:04 +0800 Huan Yang <link@vivo.com> wrote:
>
>> @@ -3652,7 +3654,10 @@ static struct mem_cgroup *mem_cgroup_alloc(struct mem_cgroup *parent)
>>   	int __maybe_unused i;
>>   	long error;
>>   
>> -	memcg = kzalloc(struct_size(memcg, nodeinfo, nr_node_ids), GFP_KERNEL);
>> +	memcg = likely(memcg_cachep) ?
>> +			kmem_cache_zalloc(memcg_cachep, GFP_KERNEL) :
>> +			kzalloc(struct_size(memcg, nodeinfo, nr_node_ids),
>> +				GFP_KERNEL);
> Why are we testing for memcg_cachep=NULL?

Due to cgroup_init ahead of initcall, hence, root_mem_cgroup create will trigger NULL pointer access.

So, here test it. :)

I don't find a more good way to avoid this.

>
>> @@ -5055,6 +5061,10 @@ static int __init mem_cgroup_init(void)
>>   		INIT_WORK(&per_cpu_ptr(&memcg_stock, cpu)->work,
>>   			  drain_local_stock);
>>   
>> +	memcg_size = struct_size_t(struct mem_cgroup, nodeinfo, nr_node_ids);
>> +	memcg_cachep = kmem_cache_create("mem_cgroup", memcg_size, 0,
>> +					 SLAB_PANIC | SLAB_HWCACHE_ALIGN, NULL);
> If it's because this allocation might have failed then let's not
> bother.  If an __init-time allocation failed, this kernel is unusable
> anyway.
>
>