[lkml]   [2019]   [Dec]   [17]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: [PATCH v6 02/10] mm/lru: replace pgdat lru_lock with lruvec lock

在 2019/12/17 上午10:16, Matthew Wilcox 写道:
>>> You still didn't fix this function. Go back and look at my comment from
>>> the last time you sent this patch set.
>> Sorry for the misunderstanding. I guess what your want is fold the patch 9th into this, is that right?
>> Any comments for the 9th patch?
> I didn't get as far as looking at the ninth patch because I saw this
> one was wrong and stopped looking. This is not the first time *with
> this patch set* that you've been told to *fix the patch*, not submit
> something that's broken and fix it in a later patch.
> I'll look at patch 9 later.

Thanks a lot for the nice cocaching and quick response!

What the problem for me here is I didn't find a bug here. From the commit_charge's explanations and mem_cgroup_commit_charge comments, as well as call path when lrucare is ture, The lock is just to guard the task migration(which would be lead to move_account) So, It's just a clean up to give up locking when !PageLRU in patch 9. And even w/o patch 9, the page just locked root_mem_cgroup's lru_lock, same as old function does, while the page isn't on any LRU. Useless, but it's still safe.

Do you mind to point out anything else I missed?

Thanks a lot!

 \ /
  Last update: 2019-12-17 14:12    [W:0.028 / U:2.924 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site