[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <513b8479-4c2d-b6c1-2081-15ea0cd0ebeb@bytedance.com>
Date: Fri, 15 Sep 2023 11:31:15 +0800
From: Chuyi Zhou <zhouchuyi@...edance.com>
To: Bixuan Cui <cuibixuan@...o.com>, Jonathan Corbet <corbet@....net>,
hannes@...xchg.org, mhocko@...nel.org, roman.gushchin@...ux.dev,
ast@...nel.org, daniel@...earbox.net, andrii@...nel.org,
muchun.song@...ux.dev
Cc: bpf@...r.kernel.org, linux-kernel@...r.kernel.org,
wuyun.abel@...edance.com, robin.lu@...edance.com
Subject: Re: [RFC PATCH v2 2/5] mm: Add policy_name to identify OOM policies
在 2023/9/15 10:28, Bixuan Cui 写道:
>
>
> 在 2023/9/14 20:50, Chuyi Zhou 写道:
>>>
>>> Delete set_oom_policy_name, it makes no sense for users to set policy
>>> names. 🙂
>>>
>>
>> There can be multiple OOM policy in the system at the same time.
>>
>> If we need to apply different OOM policies to different memcgs based
>> on different scenarios, we can use this hook(set_oom_policy_name) to
>> set name to identify which policy in invoked at that time.
>>
>> Just some thoughts.
> Well, I thought the system would only load one OOM policy(set one policy
> name), which would set the prio of all memcgs.
>
> What you mean is that multiple bpf.c may be loaded at the system, and
> each bpf.c sets the different policy for different memcgs?
No. Not multiple bpf_oompolicy.c but multiple OOM Policy in one BPF Program.
Powered by blists - more mailing lists